chg7999 m.a.sc. thesis department of chemical and ... · chg7999 m.a.sc. thesis department of...
TRANSCRIPT
CHG7999
M.A.Sc. THESIS
Department of Chemical and Biological Engineering
Lysis of Escherichia coli for the Recovery of Pentamerised Single-Domain
Antibody Used for the Gender Specific Separation of Bovine Sperm
THÈSE DE M.Sc.A.
Département de Génie Chimique et Biologique
Jordan O’Reilly
University of Ottawa
© Jordan O’Reilly, Ottawa, Canada, 2016
ii
ABSTRACT
Gender of animal offspring is of great interest to farmers where gender selection is achieved
via the separation of male-bearing from female-bearing sperms prior to performing artificial
insemination. A start-up company (Ab Biotech Inc.) has developed a technique for gender
selection based on the production of an intracellular single-domain antibody (sdAb) using the
bacterium Escherichia coli capable of sexing bovine sperm. The purpose of this research was to
provide a recommendation to Ab Biotech Inc. for the lysis of E. coli. An efficient lysis technique
was required in order to release the intracellular sdAb. In the dairy industry, sexing for female
calves is preferred since male calves are not useful for the purpose of milk production. Multiple
lysis techniques were tested in order to provide a feasible recommendation for Ab Biotech Inc.
These techniques included high pressure homogenization, sonication, bead milling and
enzymatic/chemical lysis using lysozymes and Triton X-100. Required lysis time, extent of lysis
and potential operating costs were contributing factors for determining an optimal technique. The
extent of lysis was determined by quantifying the total amount of released protein using SDS-
PAGE densitometry. It was recommended to choose bead milling for potential process upscaling
since a large amount of fractional lysis (0.70) was obtained over a short amount of lysis time (3
min) with an inexpensive ($9.50/kg) 0.3 mm mixture of glass beads.
iii
RÉSUMÉ
Le sexe de la progéniture des animaux est d'un grand intérêt pour les agriculteurs, où la
sélection du sexe est obtenue par la séparation des spermes menant à un mâle ou une femelle avant
que l'insémination artificielle ne soit effectuée. Une entreprise en démarrage (Ab Biotech Inc.) a
mis au point une technique de sélection du sexe de la race bovine basée sur la production d'un
anticorps à domaine unique (ATDU) intracellulaire par la bactérie Escherichia coli (E. coli).
L’objectif principal de cette thèse était de fournir une recommandation à Ab Biotech Inc. pour la
sélection d’une méthode pour la lyse des cellules E. coli. Une technique de lyse efficace est requise
afin de libérer un anticorps à domaine unique (ATDU) intracellulaire utilisé pour le sexage des
spermes bovins. Dans l'industrie laitière, le sexage des veaux femelles est préféré car les veaux
mâles ne sont pas utiles aux fins de la production de lait. De multiples techniques de lyse ont été
étudiées afin de fournir une recommandation la plus appropriée pour Ab Biotech Inc. Les
techniques de lyse qui furent considérées sont l'homogénéisation à haute pression, la sonication,
le broyage par billes et la lyse enzymatique/chimique avec les lysozymes et le Triton X-100. La
durée de la lyse, le pourcentage de lyse obtenue et les coûts potentiels d'exploitation ont été les
principaux facteurs considérés pour déterminer la technique de lyse la plus appropriée pour cette
application. Le pourcentage de lyse est déterminé par la quantité totale de protéines libérées qui
est obtenue par densitométrie des gels de SDS-PAGE. Il a été recommandé de choisir le broyage
par billes pour la mise à l’échelle du procédé de lyse puisque cette technique de lyse a donné le
plus grand pourcentage de lyse (70 %) obtenu pour un temps de lyse relativement court (3 min).
De plus le mélange de billes ayant un diamètre moyen de 0.3 mm est relativement peu dispendieux
(9.50 $/kg).
iv
ACKNOWLEDGEMENTS
First I would like to thank my supervisors, Dr. Jules Thibault and Dr. Albert Parisien, for
their guidance and support throughout the entire two-year process.
I would also like to thank Robert Letellier, CEO of Ab Biotech Inc., for giving me the
opportunity to work with his company to complete this thesis.
I would like to recognize Marc-André Hamelin, also from Ab Biotech Inc., who helped me
with fermentations, protein purification and western analysis.
I would like to thank Dr. Christopher Lan for his help revising my literature review.
I would like to acknowledge Sophie Boisvenue and France Charbonneau from the
Technology Access Centre in Bio-Innovation at Collège La Cité for letting me use their wonderful
facilities.
I would also like to thank Dr. Michel Gilbert and Dr. David Watson from NRC Institute
for Biological Sciences Human Health Therapeutics for access to their High Pressure
Homogenizer and help performing some of the experiments.
I would like to acknowledge the financial support from the NSERC Engage Grant and from
the University of Ottawa Admissions Scholarship.
I would like to thank my family and my colleagues from D218 who have made this
experience very enjoyable.
Finally I would like to thank my girlfriend Janika Ouimette for all of her love and support
throughout this entire process.
v
TABLE OF CONTENTS
Abstract ........................................................................................................................................... ii
Résumé ........................................................................................................................................... iii
Acknowledgements ........................................................................................................................ iv
Table of Contents ............................................................................................................................ v
List of Figures .............................................................................................................................. viii
List of Tables ................................................................................................................................. ix
Nomenclature .................................................................................................................................. x
Chapter 1: Introduction ................................................................................................................... 1
Chapter 2: Literature Review .......................................................................................................... 5
2.1 Flow cytometry ..................................................................................................................... 5
2.1.1 Methodology ................................................................................................................... 5
2.1.2 Drawbacks ...................................................................................................................... 7
2.2 Pentameric single-domain antibodies for sperm sexing ........................................................ 9
2.3 Production strain: E. coli ..................................................................................................... 10
2.3.1 Importance of E. coli lysis ............................................................................................ 11
2.3.2 The E. coli cell wall ...................................................................................................... 11
2.4 Methods of lysis .................................................................................................................. 12
2.4.1 Chemical lysis............................................................................................................... 13
2.4.2 Enzymatic lysis (lysozymes) ........................................................................................ 14
2.4.3 High pressure homogenization ..................................................................................... 16
2.4.4 Sonication ..................................................................................................................... 18
2.4.5 Bead milling ................................................................................................................. 21
2.5 Quantification of the extent of lysis .................................................................................... 24
2.5.1 Direct cellular analysis ................................................................................................. 24
2.5.2 Indirect analysis (quantification of cellular products) .................................................. 26
2.6 Conclusion ........................................................................................................................... 31
Chapter 3: Methods and Materials ................................................................................................ 33
3.1 Introduction ......................................................................................................................... 33
vi
3.2 Protein (antibody) description ............................................................................................. 33
3.3 Host organism (E. coli) ....................................................................................................... 35
3.4 E. coli production ................................................................................................................ 36
3.4.1 Fermentation performed in flasks ................................................................................. 36
3.4.2 Fermentation performed in bioreactors ........................................................................ 38
3.5 Lysis Procedures ................................................................................................................. 40
3.5.1 Methods and materials for Section 4.1 ......................................................................... 40
3.5.2 Methods and materials for Section 4.2.2 ...................................................................... 40
3.5.3 Methods and materials for Section 4.2.3 ...................................................................... 41
3.5.4 Methods and materials for Section 4.3 ......................................................................... 41
3.5.5 Methods and materials for Section 4.4 ......................................................................... 43
3.6 SDS-PAGE .......................................................................................................................... 44
3.6.1 SDS-PAGE procedure .................................................................................................. 44
3.6.2 SDS-PAGE analysis ..................................................................................................... 46
3.7 Western blot analysis .......................................................................................................... 47
Chapter 4: Evaluation of lysis techniques ..................................................................................... 51
4.1 High pressure homogenization ............................................................................................ 51
4.1.1 Results and discussion .................................................................................................. 52
4.1.2 Summary ....................................................................................................................... 54
4.2 Bead milling ........................................................................................................................ 54
4.2.1 Bead milling: effect of bead to lysate ratio ................................................................... 55
4.2.2 Bead milling: cell growth cycle point and time ............................................................ 55
4.2.3 Bead milling: effect of bead size .................................................................................. 63
4.3 Sonication ............................................................................................................................ 69
4.3.1 Results and discussion .................................................................................................. 70
4.3.2 Model development ...................................................................................................... 73
4.3.3 Summary ....................................................................................................................... 77
4.4 Synergistic lysis using lysozyme and Triton X-100............................................................ 78
4.4.1 Results and discussion .................................................................................................. 79
4.4.2 Summary ....................................................................................................................... 81
4.5 Overall Comparison ............................................................................................................ 81
Chapter 5: Conclusion and Recommendations ............................................................................. 84
vii
5.1 Conclusion ........................................................................................................................... 84
5.2 Recommendations ............................................................................................................... 84
Chapter 6: References ................................................................................................................... 86
viii
LIST OF FIGURES
Figure 1-1: Flow sheet depicting the forms of lysis presented throughout the document .............. 4
Figure 2-1: Mechanism of sperm sexing through flow cytometry ................................................. 6
Figure 2-2: A visual comparison between a monomeric sdAb and a pentameric antibody ......... 10
Figure 2-3: An illustration of the cellular wall of E. coli bacteria ................................................ 12
Figure 2-4: Lysis mechanism of peptidoglycan through the use of lysozymes ............................ 15
Figure 2-5: High pressure homogenization mechanism ............................................................... 16
Figure 2-6: Typical sonication set-up ........................................................................................... 19
Figure 2-7: Continuous bead milling schematic diagram ............................................................. 21
Figure 3-1: A typical SDS-PAGE of E. coli proteins ................................................................... 47
Figure 3-2: An example of results obtained for SDS-PAGE and western blot ............................ 49
Figure 4-1: SDS-PAGE of E. coli proteins. .................................................................................. 52
Figure 4-2: Fractional release of proteins durinf HPH experimentation ...................................... 53
Figure 4-3: E. coli TG1 growth curve for Fermentation 1 ............................................................ 57
Figure 4-4: E. coli TG1 growth curve for Fermentation 2 ............................................................ 57
Figure 4-5: E. coli TG1 growth curve for Fermentation 3 ............................................................ 58
Figure 4-6: A-C: SDS-PAGE of E. coli proteins for Fermentation 3 ........................................... 59
Figure 4-7: A comparison of the amount of fractional lysis of E. coli cells over 9 min of bead
milling with cells that were harvested from the exponential or stationary phase. ........................ 60
Figure 4-8: Effect of bead milling time on cell lysate temperature .............................................. 62
Figure 4-9: Microscopic images of glass beads used for the operation of milling E. coli ............ 64
Figure 4-10: Particle size distribution of the “0.1 mm” diameter beads ....................................... 64
Figure 4-11: Particle size distribution of the “0.3 mm” diameter bead mixture ........................... 65
Figure 4-12: Particle size distribution of the “0.5 mm” diameter beads ....................................... 65
Figure 4-13: SDS-PAGE of E. coli proteins from bead milling experiment comparing lysis as a
function of glass diameter ............................................................................................................. 66
Figure 4-14: Comparison of the amount of fractional lysis of E. coli cells over 6 min of bead
milling from the bead milling experiment comparing lysis as a function of glass bead diameter 67
Figure 4-15: The effects output control, duty cycle, optical density and time (D) on the fractional
amount of lysis during sonication ................................................................................................. 72
Figure 4-16: A plot of the predicted amount (using Equation (4.1)) of fractional lysis versus the
amount of experimental lysis during sonication ........................................................................... 74
Figure 4-17: Pareto plot for initial model of sonication experimentation .................................... 75
Figure 4-18: A plot of the predicted amount (using Equation (4.2)) ............................................ 76
Figure 4-19: SDS-PAGE of E. coli proteins from experimentation with lysozymes and Triton X-
100................................................................................................................................................. 79
Figure 4-20: A comparison of the amount of fractional lysis of E. coli cells during experimentation
with lysozymes and Triton X-100................................................................................................. 80
ix
LIST OF TABLES
Table 2-1: Approximate extinction coefficients for 1 mg/mL protein solutions .......................... 27
Table 3-1: A breakdown of the values of each parameter during the sonication experimentation
....................................................................................................................................................... 43
Table 4-1: Cellular growth phase classification for cell fermentations 1, 2 and 3. ...................... 58
Table 4-2: A breakdown of the values of each parameter during the sonication experimentation, as
well as experimental and predictive results (using Equations (4.1) and (4.2)) ............................. 71
x
NOMENCLATURE
A Absorbance
Β Beta BCA Bicinchoninic acid BSA Bovine serum albumin CFU Colony Forming Unit DNA Deoxyribonucleic acid
E Extinction coefficient E. coli Escherichia coli
EDTA Ethylenediaminetetraacetic acid FISH Fluorescence in situ hybridization HRP Horseradish Peroxidase GHCl Guanidine hydrochloride
IM Inner membrane IMAC Immobilized metal affinity chromatography IPTG Isopropyl β-D-1-thiogalactopyranoside
LB Lysogeny broth LPS Lipopolysaccharides
mAB Monoclonal antibody NAG N-acetylglucosamine NAM N-acetlymuramic acid
NTA Nitrilotriacetic acid
OD Optical Density OM Outer membrane P Pellet
PBS Phosphate buffered saline PCR Polymerase chain reaction
PL Phospholipid layer PMSF Phenylmethanesulfonylfluoride PVDF Polyvinylidene fluoride rcf Relative centrifugal force
rpm Revolutions per minute S Soluble
sdAb Single-domain antibody SDS Sodium dodecyl sulfate SDS-PAGE Sodium dodecyl sulfate - polyacrylamide gel electrophoresis SSP Sex specific protein T Total
TB Terrific Broth TBST Tris-buffered saline and Tween 20 UV Ultraviolet VhH Variable heavy fragment
1
CHAPTER 1: INTRODUCTION
The agricultural industry has been transformed in the past half century from small family
farms producing a wide variety of products to very large and increasingly specialized operations.
This is the case for dairy farms where the majority of the farm operation is oriented to milk
production. When cows are artificially inseminated, it is highly desirable to resort to gender
selection to increase the occurrence of female calves since male calves have no value, and often
negative values, for this type of operation. However, for the beef industry the opposite prevails;
male calves would be preferentially selected since the males grow faster than the females [1]
resulting in higher beef productivity. For milk production farms, resources used to take care of the
unwanted male calves during pregnancy, birth, and infancy are deemed to be wasted since these
calves cannot potentially contribute to the specified operation. Therefore, gender selection is
highly desirable due to cost and resource saving potential.
Sex predetermination in livestock, to provide the producers with faster genetic progress
and greater flexibility, has been the subject of research for many decades. The current technology
for gender pre-selection, flow cytometry, relies on the difference in the amount of DNA present in
X- and Y-sperms (X sperms create females while Y sperms create males). For cattle the DNA
content is approximately 4 % greater in X-chromosomal sperms [2]–[5]. Flow cytometry is mostly
confined to the cattle breeding industry with relatively good success for pre-sexing except that the
sexed sperm doses are too expensive, $80-$100/dose for female straws, for widespread application
[2], [5], [6]. Furthermore, sexing sperm with a flow cytometer is considered to be inefficient (42-
50 % of each sample is able to be sorted [2]). Because of these reasons, sexed sperm is normally
reserved for first pregnancy in heifers where fertility is the highest [2], [7], [8]. Therefore, there is
2
a clear need for a rapid, economic, efficient and large-scale method for separating X or Y-sperm
populations.
One technique, developed by Ab Biotech Inc., is to use specific single-domain
pentamerised antibodies that are able to capture and remove male spermatozoa from bull semen
enabling the selectivity of female-forming spermatozoa. The antibody is able to attach to antigens
specifically found only on the surface of male-forming spermatozoa providing a means of
selectively separating males from females through immobilized metal affinity chromatography
(IMAC) since the antibody exhibits a histidine-tag (His-tag).
Typically, antibodies are produced in eukaryotic cells since this type of cell contains the
cellular machinery needed to provide protein glycosylation and proper protein folding [9].
However, the antibody used in the following research is a single-domain antibody (sdAb) which
is much simpler structurally than conventional antibodies hence requiring minimal folding, and
does not require glycosylation in order to bind the target epitope [10]. Therefore, this antibody can
be produced using a system that is easier and cheaper to maintain such as prokaryotic cells. The
host organism that has been chosen for the production of the antibody in this investigation is
Escherichia coli , a unicellular gram-negative bacterium which can be grown in large amounts
easily and cheaply [11].
The major steps of the production process involve 1) a fermentation to produce high cell
concentration, 2) an induction to initiate the production of high amounts of antibodies, 3) cell lysis
to expose the antibodies, and 4) the recovery of antibodies from the solution using a targeted
separation technique such as IMAC. All steps are paramount for the success of the economic
viability of this process. However, this thesis focuses on the cell lysis step which is critical to
achieve the highest possible proportion of soluble antibodies available for the purification step. If
3
a fraction of antibodies is unable to be extracted from cellular debris, the efficiency of the entire
process decreases significantly leading to increased operating costs. The lysis step poses
significant challenges since there are two cellular walls that need to be breached in order to access
the antibodies located within E. coli cells. Another challenge consists in ensuring that the
antibodies remain stable and active upon release from the cell.
The partner company, Ab Biotech Inc., is working simultaneously on the development of
all four production steps including cell lysis at large scale. Large-scale lysis of cells cannot rely on
small-scale lysis techniques commonly used in laboratory experiments. To address this problem,
various mechanical and chemical/enzymatic lysis techniques were investigated in this thesis. A
comparative study was performed to determine which technique would yield a higher relative
protein recovery as well as determining which technique could be efficiently scaled-up for
potential large-scale operation. This was achieved by identifying which lysis techniques were
capable of achieving maximum cell lysis under laboratory-scale conditions. Mechanical lysing
techniques that were tested in this investigation include bead milling, high pressure
homogenization and sonication. Chemical/enzymatic techniques that were tested include Triton
X-100 and lysozymes. The series of experiments performed throughout this investigation and
presented in this thesis is summarized in Figure 1-1.
4
Figure 1-1: Flow sheet depicting the forms of lysis presented throughout the document.
In summary, to meet the company’s need, the objective of this research was to develop an
efficient lysis method that maximized the amount of soluble pentameric antibodies produced
through E. coli fermentations for downstream processing. However, throughout the research
performed, it was determined that the amount of the target antibody produced was minimal and
was therefore difficult to quantify. Since it was difficult to produce high concentrations of the
target antibody with the current development of the genetically-modified cellular strain, the focus
of this thesis has been expanded to include the quantification of all recovered proteins as a standard
for determining the applicability of specific methods of lysing E. coli cells.
Lysis Experimentation
Mechanical
High Pressure Homogenization
Bead Milling Sonication
Chemical/Biological
Lysozymes Triton X-100
5
CHAPTER 2: LITERATURE REVIEW
This chapter briefly presents background information on the methods that are used to separate
sperms for gender selection, the production of a specific protein for sexing and the different lysis
techniques used to liberate the cell content and make it accessible for specific targeted recovery.
This information will help to better understand the challenges that are faced in gender selection
and cell lysis.
2.1 Flow cytometry
To this day, flow cytometry is the only viable method for sexing live sperm commercially
and is only available through the company Sexing Technologies [7], [12]. The sperm separation is
based on the fact that X-sperms contain, in the case of cattle, approximately 4 % [2]–[4] more
DNA than the Y-Sperms [2], [13]–[15]. The premise behind flow cytometry is to measure the
amount of DNA content of each individual sperm and sex the sperm based upon the resulting
measurement. Currently, flow cytometry is able to correctly sex sperm with a 90 % success rate
[2], [4], [7], [8], [12], [14]. However, only about 42-50 % of the sperm in each sample are sorted
leaving behind more than half of the desired sperm [2].
2.1.1 Methodology
In flow cytometry, the sperm sample is subjected to a fluorescing dye (Hoechst 33342)
which is able diffuse through the cell membrane of the sperm and bind with the sperm’s DNA [3],
[7], [12]–[14]. Since, for cattle, X-sperms have a 4 % higher DNA content than Y-sperms, the X-
sperms will fluoresce with 4 % greater intensity than the Y-sperms [7], [12], [14], [16].
6
The sperm cells are then passed through the flow cytometer exiting the apparatus one at a
time. The outlet flowrate of sperm was optimized through the development of a beveled injection
needle which ensures the correct orientation of sperm leaving the flow cytometer at high flow rates
[3], [7], [14]. Upon exiting the apparatus, the cell sample is broken into droplets by a vibrator [14],
[17]. This process typically forms 70 000 to 80 000 drops per second [2]. Each droplet is analyzed
by a solid-state laser which detects whether the sperm is an X-sperm or a Y-sperm based upon the
level of fluorescence that the sperm emits [12], [14]. If a droplet contains an X-sperm, a positive
charge is added to the droplet whereas, if a droplet contains a Y-sperm, a negative charge is added.
Each droplet then falls between two electrically-charged plates [2], [14]. One plate emits a positive
electric field while the other plate emits a negative electric field. Depending on the charge assigned
to the droplet, the droplet will be either drawn toward the positively or the negatively charged plate
(the opposite charge to which the droplet has been assigned). The separated sperms are collected
at the bottom of each respective plate [2], [14]. Figure 2-1 provides a schematic diagram describing
the mechanism for sperm sexing via flow cytometry.
Figure 2-1: Mechanism of sperm sexing through flow cytometry [2].
7
Along with the Hoechst 3342 fluorescent dye, food colorant is also added to the cell sample
prior to being passed through the flow cytometer [3], [7], [12]. Sperm cells with damaged
membranes are not able to keep the food colorant from entering the intracellular space. Therefore,
the food colourant acts to quench the fluorescent dye lowering the fluorescence [3], [7], [12]. When
the damaged sperms are analysed by the laser, they do not meet the threshold of required
fluorescence. The damaged sperms are not given a charge and are able to pass freely between the
two electrically-charged plates and are discarded. This same principle applies to droplets which 1)
do not contain sperms (no fluorescence is observed), 2) contain multiple sperms (fluorescence is
too high), or 3) contain sperms which are indistinguishable relative to their DNA content
(abnormal amount of fluorescence). In each of these cases, no charge is assigned and the droplets
pass freely through the electrically-charged plates and are discarded. After the separation process
is complete, the extent of separation can be tested by using Polymerase Chain Reaction (PCR),
Fluorescence In Situ Hybridization (FISH) or Sort Reanalysis techniques in order to determine the
X-Sperm and Y-Sperm content in the sample [3], [13].
2.1.2 Drawbacks
During the separation process, the sperms can be damaged due to the Hoechst fluorescent
dye, the laser or the shear stress due to pressure-driven sperm flow through the apparatus [3], [7],
[8], [14], [17], [18]. However, it has been determined that the fluorescent dye has no detrimental
effects on the offspring that the sperm fosters [2], [3], [14]. It has been found that lowering the
apparatus pressure from 345 to 275 kPa (50 to 40 psi) can significantly lower the amount of
damage to sperm cells [3], [7], [8], [14], [17], [18]. The viability of sperm cells is also reduced
when they are stored for longer periods prior to separation [3], [8]. Pregnancy success rates
dropped from 42 % to 37 % when bovine sperms were stored at 19°C for 20 h prior to sorting [8].
8
It has also been found that sexed cells degenerate at a faster rate than unsexed cells [3], [19].
Therefore, cryogenic techniques have been developed in order to help preserve sperm cells prior
to insemination [3], [14], [20].
Typically, a flow cytometer is able to sort 10 x 106 sperms/h [2], [4], [14]. This rate of
separation is quite low seeing that a typical dose of frozen bull sperm used for insemination
typically contains more than 20 x 106 sperms [2], [14], [19]. Even though this is the case, for most
bulls, satisfactory fertility is seen with 10 x 106 sperms per dose and, for some bulls, high fertility
can be seen even at a low dose of 2 x 106 sperms. Therefore, the low dose of 2 x 106 sperms per
insemination is commonly used to reduce the amount of time (and valuable sperms) required to
obtain a viable sperm sample for insemination [7], [8], [12], [14]. However, even though some
bulls produce semen that exhibits high fertility, heifers are typically chosen over lactating cows
for insemination since heifers are considered to be more fertile [2], [7], [8]. Lactating cows can
still achieve pregnancy rates comparable to heifers, although somewhat lower, if techniques such
as ultrasound, which are cumbersome, are used to determine the lactating cow’s fertility window
[2], [8], [14]. Proper management of heifers through 1) nutrition 2) disease control, 3) estrus
detection, 4) proper semen handling and 5) proper insemination techniques can lead to pregnancy
rates of 56 % when using 2 x 106 sexed sperms per inseminate compared to 61 % when using 10
x 106 unsexed sperms per inseminate [2], [8], [14]. If heifers are mismanaged then the pregnancy
rates are very low [2], [7], [8].
Finally, sperm sexing through flow cytometry is considered to be very expensive ($80-
$100 per female straw). The apparatus costs over $350 000 and requires skilled operators (which
represent training costs) in order to run efficiently [2].
9
2.2 Pentameric single-domain antibodies for sperm sexing
It has been hypothesized that sperm cells can be sexed based upon sex-specific proteins
(SSPs) located on the surface of the cell [20], [21]. Antibodies could then be designed in order to
attach specifically to the identified surface protein. Affinity chromatography or magnetic bead
separation would then be able to separate the X-sperms from the Y-sperms through the antibody
attachment. An approach has been developed to assist in identifying SSPs based upon the
hypothesis that sex-specific proteins are evolutionarily more highly conserved than non-SSPs [21].
Rather than developing a monoclonal antibody (mAb) to target a SSP, a single-domain
antibody (sdAb also known as a VHH or nanobody) can be designed for this purpose. Compared
to mAb (~160 kDa), sdAbs are significantly smaller (~15 kDa) since they only contain a single
variable heavy domain. In comparison, mAbs have 2 light and 2 heavy chains containing both
variable and constant regions. The small size of sdAbs allow them to reach antigens on the surface
of a cell which would be unreachable by mAbs [22]–[24]. They are 1) very soluble, 2) resistant to
high temperatures (up to 90°C), 3) resistant to extreme pH and 4) resistant to protease digestion.
Extensive folding and glycosylation are not required to produce sdAbs. Therefore, yeasts or
bacteria (such as Escherichia coli) can be used for production of sdAbs [22], [23], [25]–[29]. This
can greatly reduce production costs compared to eukaryotic production systems.
A major drawback for sdAbs is that they can have a low affinity for surface antigens [30]–
[32]. Pentamerization of sdAbs can lead to an increase in affinity to target antigens 1 000-10 000
times greater than monomeric sdAbs [30]–[33]. The sdAbs are fused to a pentamerization protein
such as the B subunit of verotoxin (VT1B) (shiga-like toxin) in order to promote pentamerization
[24], [30], [32], [33]. The pentamerization process occurs simultaneously as the fused
10
sdAb/pentamerization protein complex is produced [30], [32], [33]. A comparison between a
monomeric sdAb and a pentameric sdAb complex is illustrated below in Figure 2-2.
Figure 2-2: A visual comparison between a monomeric sdAb (A) and a pentameric antibody (B). RBS: ribosome
binding site. OmpA: OmpA leader sequence. VHH: single-domain antibody. VT1B: verotoxin 1 subunit B. c-myc: c-
myc epitope tag. (His) 5: penta-histidine epitope tag [31].
Even though the pentamerised version is larger than the monomeric version of the sdAb, it
has been found that there is no significant change in binding site accessibility to surface antigens
[33]. The pentameric version also has 1) excellent thermostability, 2) resistance to proteases (such
as trypsin and chymotrypsin), and 3) negligible aggregation (high solubility) [33]. Finally,
pentameric sdAb complex can also be produced in bacteria (E. coli) or yeasts in very high amounts
(up to 200 mg/L) [33].
2.3 Production strain: E. coli
In this investigation, E. coli was used to produce the desired protein. In this section, some
information on E. coli is presented to gain a better understanding on the importance and challenges
of producing and recovering the targeted protein.
11
2.3.1 Importance of E. coli lysis
E. coli is a gram-negative bacterium that is an important platform for the production of
intracellular products. Many reasons justify the importance of E. coli: 1) they are well-known and
well-characterized bacteria [11]; 2) they can be grown very quickly to large cell densities [11]; and
(3) they can be produced using inexpensive media and fermentation equipment [11]. They have
been the subject of extensive studies with their genome being fully decoded and an abundance of
genetic engineering tools have been developed for these bacteria, which make them a very well-
established platform for both genetic engineering research and expression of recombinant proteins
of interest [11]. However, one of the biggest drawbacks to producing desired compounds with E.
coli is the fact that these cells do not normally secrete proteins to the environment; proteins that
are normally produced stay trapped within the confines of the cellular structure. This poses a
challenge since the cellular wall of E. coli cells now has to be disrupted in order for the intracellular
proteins to reach the surrounding environment. The following sections describe possible
mechanisms that can be utilized in order to disrupt (or permeabilize) the cellular wall of E. coli
cells in order to expose a protein of interest.
2.3.2 The E. coli cell wall
In order to release intracellular proteins, cells either need to be lysed or become permeable
to the proteins located within. As seen in Figure 2-3, the cell wall is composed of three layers: an
inner membrane and outer membrane (composed of phospholipids) and a thin peptidoglycan layer
(composed of a lattice structure of N-acetlymuramic acid (NAM) and N-acetylglucosamine (NAG)
molecules) which is situated in between the two membranes.
12
Figure 2-3: An illustration of the cellular wall of E. coli bacteria [34].
Proteins of interest can either be found within the inner membrane (cytoplasmic space) of
the E. coli cell or in the periplasmic space (space between the two membranes) if an export peptide
is fused to the recombinant protein. It can be of benefit for the protein of interest to be located
within the periplasmic space since only the outer membrane would need to be removed. Since
fewer native proteins are found in this space, purification can be considerably simplified [11], [35].
However, if the protein is not engineered to be translocated to the periplasmic space, then both
membranes and the peptidoglycan layer of the cell wall need to be lysed. The following methods
allow for the recovery of proteins that are located within the E. coli cell.
2.4 Methods of lysis
Lysis is the first step in the purification process of intracellular bio-products. There are
many variables and challenges to consider when choosing the proper form of lysis since a
technique that may be optimal for the recovery of a specific product may be unsuitable for a
different product. Therefore, it is suggested to properly analyse and test various forms of lysis to
13
ensure the highest recovery of the desired product. The following sections illustrate various lysis
techniques that can be applied to E. coli for cellular product recovery.
2.4.1 Chemical lysis
As the name implies, chemical lysis uses a chemical to lyse the cell by penetrating,
destabilizing or disintegrating the cell wall barriers [36].
Triton X-100 is a non-ionic detergent that is able to release intracellular proteins from E.
coli cells by solubilizing the phospholipids of the inner and outer membranes allowing for the
release of intracellular proteins [37]–[39]. Sarkosyl is an anionic detergent that is also able to
solubilize the inner membrane of E. coli cells [40]. However, Sarkosyl is unable to solubilize the
outer membrane of E. coli. Therefore, other compounds are also needed such as Triton X-100 to
disrupt the outer membrane for the release of intracellular proteins.
Guanidine hydrochloride (GHCl) is a chaotropic agent (disrupts hydrogen binding network
between water molecules) which has the ability to solubilize proteins from cytoplasmic membrane
fragments [38]. Studies have shown that when both Triton X-100 and GHCl have been utilized
together they can have a synergistic effect on the amount of intracellular protein released from E.
coli. Hettwer and Wang (1989) have shown that for concentrations of 0.1 M GHCl and 2 % (v/v)
Triton X-100, 50 % of both intracellular and membrane proteins were released in comparison to
almost no protein released when each of these components were used on their own at these
concentrations [38].
As stated above, Triton X-100 is able to solubilize the outer and inner membranes of E.
coli in order to promote protein release [37], [38]. However, this is only the case when there are
no divalent cations present in the lysis buffer [41]. When Mg2+ ions are present in the lysis buffer,
14
Triton X-100 is unable to solubilize the outer membrane of the cells since the magnesium ions are
able to stabilize the outer membrane from the effects of the detergent. However, EDTA (a chelating
agent) is able to remove the divalent magnesium cations enabling Triton X-100 the ability to
solubilize the outer and inner membranes of E. coli cells [41], [42]. It should be noted that using
each of these compounds on their own yields a minimum amount of protein recovery due to poor
cell lysis [43]. EDTA has no effect on the peptidoglycan layer found in E. coli cells [44]. It has
been seen that some strains of E. coli like O104:H21 have higher resistance to the effects of EDTA
than others like O157:H7 ATCC 43895 [45].
EDTA can also be used with the chaotropic agent urea to initiate E. coli cell lysis. As also
seen with the chaotropic agent (GHCl), urea is able to solubilize membrane fragments. It has been
determined that concentrations of urea ranging from 0.3 mM to 6 M have been effective in
releasing intracellular proteins from both logarithmic and stationary phase E. coli at levels
comparable to mechanical disruption [46]. However, proteins get denatured when high
concentrations of urea (2 to 8 M) are used and need to be dialysed in proper buffer to enable
refolding in the correct structure [47].
2.4.2 Enzymatic lysis (lysozymes)
Lysozymes are enzymes purified from chicken egg whites that can be used to lyse E. coli
cells for the recovery of intracellular proteins [48]. Lysozymes can hydrolyse the beta 1-4 linkages
between N-acetlymuramic acid and N-acetyl glucosamine units that make up the peptidoglycan
layer as illustrated in Figure 2-4 [44], [49]–[52]. The optimal temperature for lysozyme cleavage
has been determined to be 35°C [50]. The mechanism of peptidoglycan cleavage can be visualized
in Figure 2-4.
15
Figure 2-4: Lysis mechanism of peptidoglycan through the use of lysozymes [52].
Gram-positive bacteria are easily lysed using lysozymes since there is no membrane
surrounding their large layer of peptidoglycan [50], [53]. In comparison, gram-negative bacteria
are more resistant to lysis from lysozyme attack due to a cellular membrane encasing the
peptidoglycan layer [50]. Therefore, lysozyme on its own is unable to lyse E. coli cells at a very
high efficiency since the outer membrane has not been ruptured [50].
When using lysozymes, chemically treating E. coli cells with EDTA or Triton X-100 can
help disrupt the cellular membrane leading to an enhanced cell lysis. EDTA and Triton X-100 have
been demonstrated to have the capability of disrupting the E. coli cell’s outer membrane [50], [53],
[54] enabling lysozymes to gain access to the peptidoglycan layer, and thus causing lysis. It has
been determined that optimal conditions for an EDTA/lysozyme synergistic lysis is 100-800
µg/mL of EDTA and 25-50 µg/mL of lysozyme for cells that have been harvested during
exponential and late stationary phases [50].
16
2.4.3 High pressure homogenization
High pressure homogenization is able to lyse E. coli cells physically by applying a
significantly high shear rate to each of the cells. A cellular suspension under high pressure is forced
through a small orifice in the homogenizing unit after which it collides with an impact ring causing
cellular lysis. It has been determined that gram-negative bacteria (ex. E. coli) are more susceptible
to lysis using this technique than gram-positive bacteria [55]. It has also been determined that the
effect of high pressure on E. coli bacteria is not strain-dependent. Wuytack et al. (2009) have
determined that the E. coli mutant LMM1010, (a strain of E. coli resistant to high hydrostatic
pressure) has the same resistance to high pressure homogenization as its parental strain MG1655
[55]. A schematic of a typical high pressure homogenizer is depicted in Figure 2-5.
Figure 2-5: High pressure homogenization mechanism [56].
Variables that can be easily manipulated with the high pressure homogenizer apparatus are
1) cell suspension concentration, 2) number of passes through the homogenizer, 3) pressure exerted
on the cell suspension and 4) flow rate. It has been demonstrated in literature that lysis efficiency
is independent of the cell suspension concentration [57], [58]. Since this is the case, for continuous
17
processing, the concentration of cells can be increased (while remaining below a concentration
that would be too viscous) in order to maintain a highly efficient homogenization that minimizes
the power requirement needed.
The main variable that can be adjusted for high pressure homogenization is the pressure
exerted on the cells as they flow through the apparatus [57]. Fonseca and Cabral (2002) reported
in a study the effect of homogenization pressure on the release of a periplasmic protein. They
demonstrated that pressures between 10-25 MPa could not achieve maximum periplasmic protein
release even after multiple passes through the system. Protein release throughout the tested range
was attributed to point break of the cell envelope [57]. At 50 MPa it was noted that cells were
significantly reduced in size and the periplasmic protein release was a function of the number of
passes through the system [57]. Finally when a range of 100-150 MPa was tested, it was found
that the maximum release of the periplasmic protein was seen after only one pass [57]. This
illustrates that an increase in homogenization pressure results in an increase in cellular lysis and
periplasmic protein release.
It has been observed that an increase in the initial cellular suspension temperature and a
decrease in viscosity leads to an increase in cellular deactivation during high pressure
homogenization [59]. When the temperature is increased, the viscosity of the cell suspension
subsequently decreases. However, by keeping the viscosity constant with ethylene glycol over a
range of temperatures below 45°C, it was observed that the amount of deactivation remained
constant [59]. By decreasing suspension viscosity (by varying the concentration of ethylene
glycol), higher amounts of deactivation were achieved at constant temperatures that were below
45°C [60]. However, at initial temperatures above 45°C, it was observed that the above trends did
18
not hold since the heat energy supplied on its own was high enough to contribute to cell
deactivation [59].
Pre-treating cell suspensions prior to high pressure homogenization has been shown to
increase the amount of total lysis and protein release [61]. The pressure required for maximum
protein release decreased from 34.5 MPa to 13.8 MPa when cell suspensions were treated by either
GHCl or Triton X-100 [61].
One of the downfalls of high pressure homogenization is the micronization of cellular
debris caused by multiple passes through the system [57]. When micronization occurs, cellular
debris become so small that downstream purification becomes increasingly difficult since the
cellular products can become entrained in the debris [39].
2.4.4 Sonication
Sonication involves the disruption of E. coli cells through the use of ultrasonic waves which
cause pressure fluctuations around the cell wall leading to a destabilization of the intramolecular
cohesive forces that are needed to maintain viable cellular structure [62]. A schematic
representation of a sonication device used during laboratory-scale lysis can be seen in Figure 2-6.
19
Figure 2-6: Typical sonication set-up [63].
In some cases, lysis and protein recovery efficiencies of sonication are comparable to
synergistic lysis methods using lysozymes and EDTA [50] while in other cases, the amount of
protein recovered through sonication was determined to be greater than 22 times the amount
recovered through using enzymatic lysis [64]. It has also been determined that gram-positive
bacteria are more resistant to sonication compared to gram-negative bacteria due to the rigid
peptidoglycan structure in their cell wall [65].
The main variables that affect cell lysis and subsequent protein release include power
intensity, ultrasonic frequency, volume of the cell suspension and amount of time that the cell
suspension is exposed to the ultrasonic waves. It has been shown that providing a higher power
intensity, [64], [66]–[69] or using frequencies such as 205 kHz [66] leads to an increase in lysis
and protein release. It has also been shown that a lower volume of cell suspension increases the
amount of protein released [67], [70]. However, there can potentially be an optimal volume
depending on the product being released [64]. For the release of a hepatitis B core antigen through
sonication, an optimal cell suspension volume was determined to be 15 mL [64].
20
Finally, it was shown that protein release was proportional to sonication time [58], [64],
[67], [68]. However, based upon a culmination of the other variables, time required to achieve
maximum cellular product release can vary substantially from 20 s [65] to 90 min [71]. This is an
indication that sonication kinetics depend on the microorganism itself and are influenced by the
cell physiology and growth conditions [58]. Therefore, sonication warrants experimentation in
order to evaluate its suitability as a lysis technique. Furthermore, it has been demonstrated that
prolonged exposure to sonication can lead to temperature increase or to the formation of free
radicals both of which could cause protein damage or denaturation [64], [69]–[72]. Therefore,
sonication for the release of thermally sensitive products should be performed on ice with resting
periods implemented during the process to limit protein denaturation caused by temperature
increase [70].
The effect of ionic strength and cell concentration of the cell suspension were also
determined to have no effect on the amount of cellular lysis and protein release [67], [69]. This
provides a great benefit when lysing large amounts of cells since the concentration of the cells can
be maximized reducing the power requirements needed to perform the sonication process.
Various pre-treatments prior to sonication have been shown to promote high degrees of
lysis. For example, processing cells through a freeze-thaw cycle prior to sonication has been shown
to double the amount of protein release [71]. By enhancing the amount of lysis with a pre-
treatment, less time and less acoustic power is needed thus reducing the probability of proteins
being damaged [70]–[73].
21
2.4.5 Bead milling
Another form of mechanical lysis used to lyse E. coli cells is bead milling. For industrial
purposes, bead milling would be performed on a continuous basis where a cell suspension would
be fed into the bead milling chamber containing glass beads. While flowing within the chamber,
the suspension is mixed at high speed allowing glass beads to collide with the cells. These
collisions break the cellular wall of the E. coli cells allowing for the release of intracellular
products [69], [74], [75]. After an optimal residence time (to promote maximum lysis), cellular
debris along with the released products are passed through a filter that separates the beads from
the lysate by retaining the beads within the bead milling chamber. Figure 2-7 provides a
visualization of the continuous bead milling design [76].
Figure 2-7: Continuous bead milling schematic diagram [77].
22
There have been very few experiments involving the lysis of E. coli cells via bead milling
for the purpose of recovering intracellular products. Most bead milling experimentation has been
performed on yeast cells [78]. However, based on the limited experiments performed, there are a
few key variables that can be adjusted to affect the amount of lysis occurring when using this
approach with E. coli cells. These variables include the residence time of the cell suspension
through the bead milling apparatus, the speed of the impeller, the size of the glass beads and the
amount of time lysing the cells.
It has been shown that the amount of cell lysis and protein release was proportional to the
residence time in the bead milling apparatus [69], [74] and to the impeller speed [74]. An increase
in the impeller speed mixes the cell/bead suspension at higher rates which enables the glass beads
to collide with the cells at a higher velocity. Higher velocity collisions lead to higher amounts of
cell damage and larger amounts of protein release.
Glass bead sizes were also tested to determine if they would have significant impact on the
amount of protein release. Multiple sizes of beads were tested in the range of 0.25 mm to 1.25 mm
[69]. It was determined that the optimal bead size needed to release to highest amount of aspartase
from E. coli K-12 was 0.75 mm. It is hypothesized that smaller beads are too light and fluidize
with the slurry causing minimal impact with the cells. In regards to larger bead sizes, it is
hypothesized that they are too heavy so they are unable to reach a high enough velocity for cellular
lysis [69].
The beads to cell suspension ratio also has a significant impact on the amount of cellular
lysis. In a study of the release of aspartase from E. coli, it was seen that the amount of beads
required to achieve the maximum amount of cellular lysis was 70 % of the total volume of the
chamber [69]. For E. coli lysing experiments where a bead volume of less than 70 % was used, it
23
was determined that an insufficient number of collisions between glass beads and cells occurred,
preventing maximum lysis. For experimentation above a volume of 70 %, the impeller was unable
to move the slurry at high enough speeds to ensure high enough velocities required for maximum
lysis [69].
The cellular concentration of cell suspensions has been determined to have no effect on the
efficiency of protein release from E. coli cells during lysis through bead milling [69]. This was
seen during the release of aspartase from E. coli where varying cell concentrations from 50 to 250
mg/mL had no impact on cell lysis [69]. This is an important characteristic of the system since
there is no change in efficiency as the cell concentration is increased, increasing the cellular
concentration will save on power needed to lyse cells and reduce the volume of the outlet stream
reducing downstream costs.
It was also determined that under some conditions of bead milling, the protein of interest
may end up remaining in the insoluble phase after centrifugation instead of ending up in the soluble
phase where it can be easily separated [75]. In the experiment where this was documented, the size
of glass beads that were used ranged between 0.3-0.4 mm and the lysis was performed for 20 min
[75]. In comparison, high pressure homogenization of the same cells resulted in the protein of
interest to be found in the soluble phase instead [75]. There can be two explanations for this. The
first explanation would be that the conditions that were used for bead milling were not adequate
to properly release the protein of interest from the cells. The second explanation could be that the
protein of interest attaches to the membrane of the cell. High pressure homogenization is known
to micronize cellular debris, thus being able to remove the protein of interest from its attachment
to the cell wall [57]. Therefore, if one is trying to isolate a membrane associated protein the use of
bead milling may not be used effectively in this case.
24
Pre-treatments can also be utilized to increase the amount of lysis seen in bead milling of
E. coli. Freezing and thawing the E. coli cells prior to lysis through bead milling leads to an
increase in the amount of cellular lysis and protein release [79]. Two freezing temperatures for E.
coli pre-treatment have been tested: -4°C and -20°C. It was determined that the damage incurred
by cells that were frozen at -4°C was greater than the damage incurred at -20°C [79]. Ice crystals
are able to form within the E. coli cells under higher freezing temperatures since the cells (and
surrounding buffer) freeze at a slower rate. This formation of ice crystals leads to cellular damage
enabling a higher amount of cellular lysis after bead milling compared to cells that have not been
pre-treated through freezing [79]. Due to increased efficiency through pre-treatment, energy can
then be saved during the bead milling process.
2.5 Quantification of the extent of lysis
In this section, various techniques used for the quantification of the extent of cellular lysis
are described. The various quantification techniques can be categorized into two distinct forms:
direct cellular lysis and indirect cellular lysis. Direct cellular lysis quantification involves
observing and quantifying the amount of cellular lysis by performing techniques that are directly
related to the cell itself. In contrast, indirect cellular lysis quantification depends upon the
quantification of intracellular material such as proteins in order to indirectly determine the amount
of cellular lysis that has occurred during the experiment.
2.5.1 Direct cellular analysis
This section briefly describes the lysis quantification techniques that examine the viability
of cells following lysis experiments.
25
2.5.1.1 Microscopic cell count
One method to determine the extent of lysis that has occurred during a lysis experiment is
to observe the cell lysate sample under a microscope. A phase-contrast microscope can be used to
determine whether or not the cellular structure of individual cells has been compromised. By
measuring and comparing the ratio of compromised cells to surviving cells in a lysed sample with
the same ratio prevailing in an original sample, the extent of lysis achieved during an experiment
can be determined. However, this method can be time-consuming and cumbersome [65].
2.5.1.2 Colony forming units (CFUs)
Another method to directly determine the extent of lysis consists in performing a plate
count. Samples are withdrawn from the cellular lysate, serially diluted in buffer then spread on
agar plates in order to form colonies. Each colony produced is considered to be a colony forming
unit (CFU) which can be directly linked to the amount of lysis that has occurred during the
experiment. A lower amount of CFUs indicates a larger extent of lysis and, as a result, a more
efficient lysis method. This method is considered time consuming since agar plates must be
prepared and inoculated plates must incubate a relatively long time to allow for the growth of the
CFUs [65].
2.5.1.3 Optical density at 600 nm
It has been determined that the extent of lysis of E. coli cells can be determined by the
optical density of the suspension at 600 nm. Therefore, one can determine the extent of lysis by
measuring the OD of the cellular suspension at 600 nm prior to the lysis experiment and compare
it to the OD (at 600 nm) of the cellular suspension after the lysis experiment. This method can
provide a direct quantitative analysis to determine the concentration of viable cells remaining after
lysis.[65].
26
2.5.2 Indirect analysis (quantification of cellular products)
The indirect forms of cellular lysis quantification are based on the determination of the
amount of cellular products that are released and separated from E. coli cellular constituents.
Typically, the quantification is performed on the soluble phase of a supernatant after centrifuging
the lysed samples. The amount of released cellular products is then compared to the amount in the
soluble phase of the same E. coli sample prior to lysis. The difference in protein content between
the lysed and unlysed samples provides the extent of the lysis that has occurred. By using these
methods of quantification it is assumed that the amount of released cellular products is proportional
to the amount of cellular lysis that has occurred.
2.5.2.1 Biuret reaction
One method to measure the amount of proteins that has been released from E. coli is the
Biuret reaction where a strong-alkaline copper reagent is reacted with the peptide chain of the
protein to produce a purple color. The absorbance measurement is performed at 310 nm, i.e. at the
near-ultraviolet wavelength. This method is not widely used because of its low sensitivity. To
obtain an acceptable quantification, many milligrams of a sample are required [80], [81].
2.5.2.2 UV absorbance of protein at 280 nm
Measuring the ultraviolet absorbance of a protein sample is another way to quantify the
amount of proteins in a sample. At 280 nm, proteins absorb in the UV range because of their
tyrosine and tryptophan residues. For a sample containing a mixture of proteins, this method can
only provide a rough estimate of protein concentration since the amount of tyrosine and tryptophan
is variable from one protein to another. This affects the extinction coefficient that is used to
calculate the amount of proteins. For tyrosine and tryptophan-enriched proteins, the extinction
coefficient can be as high as 2.65 while samples with proteins with no tyrosine or tryptophan would
27
have extinction coefficients of 0. Therefore, it is best to use this technique to quantify purified
proteins of known extinction coefficients. Care must be taken as well to ensure that the solvent
does not interfere with the absorbance [81], [82].
2.5.2.3 UV absorbance at 205 nm
In comparison to taking the absorbance of a protein sample at 280 nm, taking the absorption
of the sample in the far-UV range (192-220 nm) can be considered to be more sensitive because it
is much less affected by the amino acid composition of proteins. Table 2-1 provides an example
of the approximate extinction coefficient for 1 mg /mL protein solutions in the range of 192-220
nm.
Table 2-1: Approximate extinction coefficients for 1 mg/mL protein solutions over the 192-220 nm range [81].
Wavelength (nm) Extinction Coefficient
220 11
215 15
210 20
206 29
205 31
200 45
192 60
It has been determined that the most consistent analysis is obtained for absorption
performed at 205 nm wavelength with extinction coefficients in the range of 28.5-33. Equation
(2.1) provides a correlation to calculate the extinction coefficient at 205 nm with higher accuracy.
The absorbance of the protein solution at 280 nm is also required in order to correct for the tyrosine
28
and tryptophan content. This equation allows for the extinction coefficient at 205 nm to be
predicted with an accuracy of less than 2 % error [83].
𝐸205
1 𝑚𝑔/𝑚𝐿= 27.0 + 120 ∗
𝐴280
𝐴205 (2.1)
The absorbance of the supernatant of a lysate sample can be measured before and after
lysis to determine the extent of lysis based on the increase in protein content found in the
supernatant (soluble) phase [81].
2.5.2.4 Absorbance of proteins at 595 nm (using Bradford Reagent)
For this method, Coomassie Blue G-250 is dissolved in perchloric acid leading to a red-
brown colour (Bradford Reagent). When a protein sample is added to this solution, the colour of
the solution turns blue and the absorbance measurement for the quantification of proteins is
performed at a wavelength of 595 nm. This method of protein quantification has been found to be
sensitive, fast and accurate. The extent of lysis can be determined by quantifying the amount of
proteins found in the supernatant of a lysate sample taken before and after lysis [81], [84]–[86].
2.5.2.5 Absorbance at 725 nm using Biuret Reagent and Folin Reagent (Lowry Method)
Another method to quantify proteins is the Lowry method which uses two types of
reactions. The first reaction uses the same copper reaction as in the Biuret method, while the second
reaction relies on the Folin-Ciocalteau reagent. The Folin-Ciocalteau reagent reacts with the
phenols in proteins such as tyrosine. Together, the two reactions create a strong dark blue color
when they are contacted with proteins. The sensitivity of this method is quite good providing a
measureable colour with only 0.1 mg/mL of proteins. This method allows for the dilution of protein
29
samples which ensure that some components that are required for protein purification do not
interfere with the quantification analysis of the proteins [81], [87]–[89].
2.5.2.6 Bicinchoninic acid (BCA) assay
The bicinchoninic acid (BCA) assay is used to quantify protein samples through the
conversion of Cu2+ to Cu+ [90], [91]. Cu2+ can be reduced by either cysteine, cystine, tryptophan,
tyrosine or the peptide bond [90]. BCA is then reacted with the resulting Cu+ in order to form an
intense purple colour who absorbance can be quantified at 562 nm [91]. If the reaction is performed
at 37°C, tryptophan, tyrosine and peptide bonds will not be fully oxidized [90]. Raising the
temperature of the reaction to 60°C increases the oxidization of these three groups which increases
the overall sensitivity of this method [90]. The BCA assay is not affected by detergents and
denaturing agents such as urea and guanidinium chloride. However, this assay is sensitive to the
presence of reducing sugars [91].
2.5.2.7 Enzymatic assays
A very popular method to determine the efficiency of lysis of E. coli cells for the recovery
of intracellular products is enzymatic assays [92]. An intracellular enzyme is selected and its
activity is quantified prior to and after a lysis experiment. An increase in enzymatic activity is
considered to be proportional to the extent of lysis achieved during the lysis experiment. Care must
be taken to ensure an excess of substrate in order to avoid the enzymatic activity plateauing at a
low activity level. In addition, the solvent must not contain any enzymatic inhibitors that could
possibly reduce the activity level of the enzyme.
30
2.5.2.8 SDS-PAGE quantification
In the SDS-PAGE (Sodium dodecyl sulfate - Polyacrylamide gel electrophoresis)
quantification method, protein samples are run on a polyacrylamide gel subjected to an electrical
voltage. Proteins are separated vertically along the gel based on their molecular weight [93].
Proteins with lower molecular weights migrate faster through the gel and are found at the lower
part of the polyacrylamide gel whereas larger proteins remain near the top of the gel [93]. After
migration, the gels are stained with either silver or Coomassie blue type stain in order to visualize
the proteins contained within the gel. Once the staining process is completed, the amount of
proteins can be quantified based on the intensity of each protein band using computer software
such as ImageJ [94].
To determine the amount of total lysis that has occurred throughout an experiment, a
sample containing the total amount of proteins in a lysate is run on the gel as well as a sample of
the supernatant of the lysate (soluble phase). The extent of lysis is determined by calculating the
ratio of the amount of proteins found in the soluble phase to the amount of proteins found in the
total lysate sample. The greater the amount of proteins released to the soluble phase, the greater
the amount of cellular lysis.
2.5.2.9 Western Blot Analysis
Western blot analysis can also be used to quantify the amount of lysis. Western blots are
utilized to determine the presence of a specific protein of interest [95]. The protein samples are
run on the same type of gel used for SDS-PAGE. The migrated proteins are then transferred to a
membrane where an antibody is added which binds to an epitope on the protein of interest. The
antibody is conjugated to an enzyme that is able to be visualized with the addition of a substrate.
Through this process, the protein of interest is able to be visualized [95]. A sample containing the
31
total amount of proteins in a lysate is run on the gel as well as a sample of the supernatant of the
lysate (soluble phase). The extent of lysis is determined by calculating the ratio of the amount of
proteins found in the soluble phase to the amount of proteins found in the total lysate sample. The
greater the amount of proteins released to the soluble phase, the greater the amount of cellular
lysis. However, for this method, the protein of interest would need to be an intracellular protein in
order to be able to indirectly determine the amount of total lysis. If the protein of interest only
localizes in the periplasmic space of the E. coli cells, then the amount of cellular lysis that could
be comfortably stated could only be the amount of periplasmic lysis.
2.6 Conclusion
Sexing of sperm through the use of antibodies can possibly become a viable alternative to
the technique currently used (flow cytometry). Using pentamerised single-domain antibodies
(sdAbs) for this purpose provides a protein that has 1) high avidity, 2) high solubility, 3) high
thermal stability, 4) resistance to proteases and 4) high production in prokaryotes. If the
pentamerised sdAb is not able to be secreted to the extracellular surroundings, then a method of
lysis is required for its release and recovery.
Many forms of lysis for the purpose of intracellular protein release have been discussed.
These methods have included 1) chemical forms of lysis (Triton X-100, EDTA, GHCl etc.) 2)
enzymatic forms of lysis (lysozymes) and finally 3) mechanical forms of lysis (high pressure
homogenization, sonication and bead milling). It was seen that for each mechanism of lysis there
were multiple variables that could be altered in order to achieve a greater amount of protein release
from E. coli cells. However, it should be noted that optimal conditions for the release of protein A
may not be optimal for the release of protein B. The same can be said when lysing different strains
32
of E. coli cells. Therefore, each system must be tested and optimized on its own in order to
determine its ideal set of conditions for each specific protein release.
One must also consider all costs associated with each lysis mechanism in order to
determine a suitable method. Costs to release specific proteins will vary depending on the system
and how efficient each method is at releasing the protein of interest. Synergistic methods of lysis
such as the chemical pre-treatment of cells prior to mechanical disruption should be considered in
order to possibly decrease the costs of expensive chemicals and lower the costs of energy required
for mechanical disruption.
Many techniques used to quantify the extent of E. coli lysis were also presented. A
thorough analysis of the system is required as well in order to choose the correct method for
quantification.
In conclusion, it can be said that the lysis of E. coli cells is a complex process with high
variability from one system to the next. Each system must be examined thoroughly in order to
choose the correct method of lysis to maximize protein release.
33
CHAPTER 3: METHODS AND MATERIALS
3.1 Introduction
This chapter outlines methods and materials that were used for all experiments performed
during this investigation. The research presented in this thesis was performed in close collaboration
with a local start-up company, Ab Biotech Inc., for the production of a protein that is targeted for
sex selection of bovine calves, through the use of E. coli cells. The specific objective of this
research project was to study the lysis of E. coli cells in order to enhance targeted protein release,
a critical step prior to recovery in the downstream process. The fact that Ab Biotech Inc. was
working on improving simultaneously several areas of the protein production resulted in a dynamic
research environment to which experiments had to be adapted. The various areas upon which work
was performed by the company were: 1) the development of high cell density fermentations to
produce high concentrations of cellular antibodies, 2) the induction of cells to express high
concentrations of cellular antibodies 3) cell lysis to release the targeted antibodies, and 4) the
recovery of the antibodies from the lysate (using immobilized metal affinity chromatography
(IMAC)). As a result, experiments needed to accommodate breakthroughs or setbacks occurring
in the individual research sections. For example, if some changes were made to improve an
upstream process, such as biomass production, the subsequent downstream processes, which
included cell lysis for protein recovery were affected. Therefore, the composition of the resulting
cellular material being processed in the cell lysis step was highly variable.
3.2 Protein (antibody) description
The proteins produced in this investigation with the E. coli cells were single-domain
antibodies which have been designed by Ab Biotech Inc. Single-domain antibodies differ from
prototypical antibodies where only a single antibody domain is present as opposed to the two heavy
34
and two light chains found with prototypical antibodies [22]–[24]. The purpose of the produced
antibody was to separate male bovine sperms from female bovine sperms for gender selection of
offspring. The protein binds specifically to antigens located only on the surface of male-forming
sperms. In order to prepare a sexed sperm sample, the antibody is mixed with semen for a period
of time that enables the male antigen/antibody binding. Nitrilotriacetic acid (NTA) magnetic
nanoparticles are then added allowing for the separation of the male sperms through the
coordination bond [96] between a poly-histidine tag present on the antibody and the NTA on the
magnetic nanoparticles. The application of a magnetic field, such as a magnet, attracts the magnetic
nanoparticles-antibody-male sperm complex. The resulting collected supernatant is now enriched
in female sperms.
Two antibodies were used throughout this investigation: 14-2M and 2-3P. Both of these
proteins are single-domain antibodies binding male sperms but having slightly different amino
acid sequences. Furthermore, these two antibodies differ in their physical form: 14-2M is a
monomer while 2-3P is a pentamer. Pentamers are actually five monomer units linked together via
a verotoxin sub-unit. A visual comparison between a monomeric and pentameric antibody is
schematically presented in Figure 2-2.
It has been shown in previous studies with other sdAbs that pentamerised antibodies have
target antigen affinity 1 000-10 000 times greater than monomeric sdAbs [30]–[33]. Therefore,
utilization of pentamers over monomers is deemed to be advantageous. However, it should be
noted that the monomeric form localizes within the periplasmic space of the E. coli cell while the
2-3P form remains in the cytoplasmic space. The localization of the monomeric form in the
periplasmic space allows for easier protein release since only one of the two cellular membranes
needs to be lysed. There are some lysis techniques, such as osmotic shock lysis, which allow for
35
selective lysing of only the outer cell membrane in order to selectively release the proteins from
the periplasmic space [11], [39].
3.3 Host organism (E. coli)
The system used by Ab Biotech Inc. for the production of their recombinant proteins relies
on the gram-negative bacterium E. coli. E. coli bacteria have been chosen for the production in
this case since they have the ability to be grown at high cell densities, enabling high levels of
recombinant protein expression using low-cost media [11]. However, the major challenge in using
E. coli is the fact that they cannot excrete recombinant proteins easily [11].
Excretion of recombinant proteins is highly desirable since the proteins are completely
removed from the bacterium and expelled into the growth medium. Translocation of the
recombinant protein into the periplasmic space of the cell can also be desirable for protein
recovery. The periplasmic space contains relatively few native proteins which facilitates
purification of the target protein. Furthermore, the release of the protein of interest from the
periplasmic space only requires the outer cell membrane to be disrupted [11]. However, since the
protein of interest is neither translocated nor excreted, both cellular membranes need to be
disrupted in order to release our protein of interest.
The E. coli TG1 (Lucigen, Middleton, WI) expression system was used in this study. E.
coli stocks (25 % glycerol v/v) (Fisher Scientific, Fairlawn, NJ) were prepared at an optical density
of 1.5 (measured using the optical density at λ = 600 nm (OD600)) from a colony of E. coli isolated
on LB Miller agar (10 g/L tryptone, 5 g/L yeast extract and 10 g/L NaCl, and 15 g/L agar) (Fisher
Scientific, Fairlawn, NJ) and frozen at -20°C.
36
3.4 E. coli production
Fermentations of E. coli cells were conducted in flasks, for the ability to conduct numerous
experiments simultaneously and in laboratory-scale bioreactors where continuous control of
operating variables such as pH, temperature control, and dissolved oxygen was possible.
Therefore, a variety of both of the aforementioned methods were used throughout the course of
the studies. It is important to note that E. coli bacteria prepared for experiments outlined in Sections
3.5.1 and 3.5.3 to 3.5.5 were produced by the staff at Ab Biotech Inc. The bacteria prepared for
the experiment outlined in Section 3.5.2 was produced by the author with the aid of Ab Biotech
Inc. staff.
3.4.1 Fermentation performed in flasks
Pre-cultures for fermentation conducted in flasks were performed in 250-mL flasks
containing 50 mL of Studier ZYM-505 medium [97], with the addition of 50 µL of 100 mg/mL
aqueous ampicillin solution (Fisher Scientific, Fairlawn, NJ) , 100 µL of glycerol (Fisher
Scientific, Fairlawn, NJ), and 100 µL of the -20 °C E. coli stock solution. A 1-L Studier ZYM-505
medium contained: 986 ml H2O, 10 g tryptone, 5 g yeast extract, 3.55 g Na2HPO4 (Sigma-Aldrich,
St. Louis, MI), 3.4 g KH2PO4 (Fisher Scientific, Fairlawn, NJ), 2.67 g NH4Cl (Fisher Scientific,
Fairlawn, NJ), 0.71 g Na2SO4, (Fisher Scientific, Fairlawn, NJ), 2.5 g glycerol, 0.25 g glucose
(Fisher Scientific, Fairlawn, NJ) and 200 µL of trace metal solution (50 mM FeCl3 (Fisher
Scientific, Fairlawn, NJ), 20 mM CaCl2 (Fisher Scientific, Fairlawn, NJ), 10 mM MnCl2 (Fisher
Scientific, Fairlawn, NJ), 10 mM ZnSO4 (Fisher Scientific, Fairlawn, NJ), 2 mM CoCl2 (Sigma-
Aldrich, St. Louis, MI), 2 mM CuCl2 (Alfa Aesar, Ward Hill, MA), 2 mM NiSO4 (Fisher Scientific,
Fairlawn, NJ), 2 mM Na2MoO4 (Sigma-Aldrich, St. Louis, MI), 2 mM H3BO3 (Fisher Scientific,
Fairlawn, NJ) and 60 mM HCl (EMD, Gibbstown, NJ)). The pre-culture was grown overnight at
37
a temperature of 30°C on an orbital shaker rotating at 300 rpm (Thermo Scientific MaxQ 5000).
After the overnight growth, the optical density was measured at 600 nm (OD600).
Based on the pre-culture optical density (OD600), a specific volume of this pre-culture was
added to 1 L of buffered modified TB broth to achieve a final OD600 value of 0.1. A 1-L of modified
TB broth contained 20 g of LB Miller (Fisher Scientific, Fairlawn, NJ), 2 g of tryptone, 19 g of
yeast extract, 2.3 g of KH2PO4 3H2O, 16.1 g of K2HPO4 (Fisher Scientific, Fairlawn, NJ), 8 mL of
50 % glycerol, 2 mL of MgSO4 (Fisher Scientific, Fairlawn, NJ) solution (30 g/50 mL), 200 µL of
trace metal solution (as described above), 1 mL of 100 mg/mL ampicillin sodium salt solution and
the remaining volume required was made up of water. All solutions were autoclaved prior to the
addition of the pre-culture except for the ampicillin solution which was sterilized using a 0.45 m
filter. The broth was then split into 500-mL flasks, each one containing 250 mL of broth. In order
to reduce the amount of foam produced during the fermentation, 100 µL of an aqueous antifoam
solution (50 % v/v of antifoam 204 (Sigma-Aldrich, Oakville, ON)) was then added to each flask.
Each flask was monitored for growth and induction was performed at an OD600 of 0.6 by
adding 250 µL of 1 M isopropyl β-D-1-thiogalactopyranoside (IPTG) aqueous solution (Fisher
Scientific, Fairlawn, NJ). The induced cultures were then left for 48 h in an orbital shaker (Thermo
Scientific MaxQ 5000) at 30 °C and 300 rpm. The optical density of the cells was monitored as
well as the pH by extracting samples from each flask 18 h and 48 h post induction.
After 48 h of growth, the broth from each flask was centrifuged at 2273 rcf (3500 rpm) for
30-45 min at 14°C (Hermle Labortechnik Gmbh Z400K) in order to separate the cells (pellet after
centrifugation) from the broth (supernatant after centrifugation). The supernatant was discarded
and each pellet was re-suspended in 25 mL of lysis buffer (6.06 g/L Tris Base (Fisher Scientific,
Fairlawn, NJ), 1.46 g/L NaCl, pH 7.2). Phenylmethanesulfonylfluoride (PMSF) (BioBasic Inc.,
38
Markham, ON) solution, a serine protease inhibitor, was added to each re-suspended pellet to a
final concentration of 1 mM and, each cell suspension was stored at -20°C until needed for an
experiment.
3.4.2 Fermentation performed in bioreactors
3.4.2.1 Batch production
The production of E. coli cells in bioreactors was carried out in a 1-L Applikon bioreactor
(Foster City, CA) with a working volume of 750 mL. The pre-culture of E. coli was prepared by
inoculating 25 mL of LB Miller medium with 100 µL of thawed glycerol stock and placed on an
orbital shaker at 37°C for 18 h to serve as an inoculum for the bioreactor. Bioreactors containing
750 mL of modified TB medium were inoculated with the pre-culture in order to obtain an optical
density of OD600 = 0.1. A 1-L solution of modified TB was described in Section 3.4.1 with the
following modifications: 1) 12 g of tryptone, 2) 24 g of yeast extract, 3) 10 g of NaCl, 4) 15 g of
Bacto-Agar (Fisher Scientific, Fairlawn, NJ), and 5) the exclusion of LB Miller.
The culture conditions were controlled using the Applikon MyControl system. The
temperature of the culture medium was kept at a constant temperature of 28°C, or as otherwise
mentioned in the text, using a heating blanket and a cooling coil. The pH of the culture medium
was monitored with an electrode (Applisense, Applikon) and kept constant at a pH 6.8 through the
addition of 2 M sulfuric acid (Fisher Scientific, Fairlawn, NJ) to lower the pH or through the
addition of 2 M NaOH (Fisher Scientific, Fairlawn, NJ) to raise the pH. The amount of dissolved
oxygen in the medium was monitored using a polarographic electrode (Applisense) and kept
constant at 30 % air saturation using filter sterilized compressed air (1 vvm) through a diffuser.
When the flow rate of air was not enough to satisfy the oxygen demand of the culture, the controller
increased the stirrer speed (2 Rushton-type blades). The stirring speed, initially set to 300 rpm at
39
the beginning of the fermentation, varied up to 1500 rpm throughout the fermentation process.
Foaming was controlled by manual addition of 50 % v/v aqueous antifoam 204 solution.
Induction was performed 17 h into the fermentation with the addition of 10X TB induction
medium in order to promote recombinant protein expression. A 1-L of 10X TB induction medium
contained 120 g of tryptone, 240 g of yeast extract, 40 g of glycerol and 1 mL of 1 M IPTG solution.
The fermentation was then allowed to proceed for an additional 24 h.
The E. coli cells were harvested by centrifuging the entire bioreactor contents in a
centrifuge (Avanti J-26 XPI, Beckman Coulter) at 1342 rcf and temperature of 4°C for 1 h to pellet
the bacteria. The pelleted bacteria were then separated into 5, 10 or 20 g samples in Falcon tubes
and frozen at -20°C until needed for experimentation.
3.4.2.2 Fed-batch fermentations
Fed-batch fermentations were performed in the same manner as the batch fermentations
outlined in Section 3.4.2.1. However, for fed-batch fermentations, a batch culture containing 400
mL of modified TB medium was grown for 20.5 h prior to induction. At 3.5 h before induction,
300 mL of aqueous feed containing 150 g/L of glucose and 15 g/L of glycerol was added
continuously over a span of 3 h to achieve a final working volume of 700 mL. At induction, 10X
TB induction media (as described in Section 3.4.2.1) was added until a final concentration of 1X
was achieved, in order to promote recombinant protein expression. E. coli cells were harvested
20.5 h after induction as described in Section 3.4.2.1.
40
3.5 Lysis Procedures
3.5.1 Methods and materials for Section 4.1
E. coli cell cultures were grown, induced and stored as specified in Section 3.4.2.1. To
initiate a lysis experiment, cells were thawed and suspended in lysis buffer (10 mM HEPES, 29.22
g/L NaCl, pH 7.4) to a final concentration of either 10 % or 20 % (w/v). During lysis experiments,
50 mL cell samples of 10 and 20 % (w/v) E. coli cell suspensions were fed through the Avestin
Emulsiflex C5 homogenizer at a pressure of 152 MPa (22 000 psi) with a flowrate of 7 mL/min .
Each sample was processed through the homogenizer for a total of 3 passes. After each pass, a 500
μL sample of the cell suspension was removed for analysis. To quantify the total protein content
in the processed sample, aliquots of each sample were taken and prepared for SDS-PAGE as
described in Section 3.6.1.
3.5.2 Methods and materials for Section 4.2.2
Three E. coli cell fermentations were performed in 1-L Applikon bioreactors (working
volume of 700-750 mL). Fermentation 1 was performed in batch mode as described in Section
3.4.2.1. Fermentation 2 was performed in fed-batch mode as described in Section 3.4.2.2. Finally
Fermentation 3 was conducted in fed-batch mode as described in Section 3.4.2.2 with the following
modifications: 1) a pre-induction temperature of 37°C and 2) a pre-induction batch phase which
lasted 16 h. To shorten the pre-induction step, the temperature was increased thereby reducing the
overall fermentation time. After the fermentations were completed, cells were harvested and stored
as described in Section 3.4.2.
When a lysis experiment was performed, frozen cell samples were removed, thawed, and
suspended in a lysis buffer (10 mM HEPES, 29.22 g/L NaCl, pH 7.4) to achieve 20 % (v/v). EDTA
was added to each sample to a final concentration of 3 mM. Each sample was milled using a
41
BioSpec BeadBeater in the 90-mL chamber with a cell suspension:glass bead (BioSpec 0.1 mm)
volume ratio of 1:1. A polycarbonate cooling jacket surrounding the milling chamber was filled
with crushed ice in order to prevent overheating of the lysate during experimentation. The BioSpec
BeadBeater was operated in a discontinuous cyclic fashion: 1 min on and 1 min off to minimize
the increase in temperature. The ice in the cooling jacket was replaced as needed. A lysate sample
of 500 µL was taken at 0, 3, 6 and 9 min of total milling time. To analyze the total protein content
in the processed sample, aliquots of each sample were taken and prepared for SDS-PAGE (as
described in Section 3.6.1).
3.5.3 Methods and materials for Section 4.2.3
E. coli cell fermentations were conducted, induced and stored as specified in Section
3.4.2.2. with the following modifications: 1) cells were grown in a 7.5-L bioreactor (working
volume of 5 L), 2) an initial batch volume of 4 L with 1 L of induction medium added at induction
time, 3) a pre-induction temperature of 32°C, 4) a pre-induction batch phase which lasted 9.5 h, 5)
3 h prior to induction, feed solution (150 g/L glucose, 15 g/L glycerol) added continuously for 3 h
and 6) a post induction phase of 12 h. Cells were lysed as described in Section 3.5.2. A lysate
sample of 500 µL was taken at 0, 3, and 6 min of total milling time. To analyze the total protein
content in the processed sample, aliquots of each sample were taken and prepared for SDS-PAGE
electrophoresis (as described in Section 3.6.1).
3.5.4 Methods and materials for Section 4.3
E. coli cell cultures were grown, induced and stored as specified in Section 3.4.2.1. The
day of the sonication experiments, cell pellets were thawed and suspended in lysis buffer (6.06 g/L
Tris Base, 1.46 g/L NaCl, pH 7.2) to achieve either a 20, 40 or 60 OD cell suspension
corresponding to 10 g/L, 20 g/L and 30 g/L dry weight of E. coli cells (where 1 OD unit ≈ 0.5 g/L
42
[98]). The E. coli cells were then lysed using the Branson Sonifier 450 sonicator. During the lysis
experimentation with the sonicator, the 50 mL Falcon tube containing 25 mL of the lysate sample
was kept in a mixture of ice and water; the ice-water mixture was replaced as needed throughout
the experiment. Because of the amount of heat produced by sonication, experiments were
performed for the desired time of lysis with a repetitive cycle of 1 minute of sonication followed
with 30 s of rest to allow the lysate temperature to decrease. Therefore, a 3 min lysis experiment
would consist of 3 1-min lysis cycles with 30 s of rest in between.
Four parameters were tested throughout the experiment: (1) the output power control was
used over a range of 2 to 6 (on a scale of 10 where 10 corresponds to the full power of 400 W), (2)
the on-off duty cycle (percentage of time the sonicator was on during the experiment) ranged from
20 to 100 %, (3) the optical density (OD) of the cell mixture was varied between 20 and 60, and
(5) the total lysis time period was varied between 1 to 5 min. A total of 17 combinations of the
four parameters were tested in order to determine the effect of each parameter on the lysis of E.
coli while minimizing the number of runs. The values of each the four parameters for each run
have been tabulated in Table 3-1. Each run was performed three times in order to produce triplicate
runs.
43
Table 3-1: A breakdown of the values of each parameter during the sonication experimentation: A represents Output
Control, B represents Duty Cycle, C represents OD and D represents Time (min).
Run A B C D
1 2 20 20 1
2 2 20 20 5
3 2 20 60 1
4 2 20 60 5
5 2 100 20 1
6 2 100 20 5
7 2 100 60 1
8 2 100 60 5
9 6 20 20 1
10 6 20 20 5
11 6 20 60 1
12 6 20 60 5
13 6 100 20 1
14 6 100 20 5
15 6 100 60 1
16 6 100 60 5
17 4 60 40 3
To analyze the total protein content in the processed lysate, 500 µL aliquots of each lysate
were taken and prepared for SDS-PAGE electrophoresis (as described in Section 3.6.1).
3.5.5 Methods and materials for Section 4.4
E. coli cell cultures were grown, induced and stored as specified in Section 3.5.3. To
perform a lysis experiment, the cells were then thawed and suspended in lysis buffer (10 mM
HEPES, 29.22 g/L NaCl, pH 7.4) to a final concentration of 20 % (w/v). Cells were then either
lysed solely using lysozymes (1350 U per g of wet cell paste) or lysed using a combination of
44
lysozymes (1350 U per g of wet cell paste) and Triton X-100 (1 % v/v final concentration). Cell
lysates were then incubated for 30-60 min on a rotary shaker at 250 RPM and room temperature
(22°C) until the suspension became viscous (an indication of cell lysis due to the release of DNA).
To lower the lysate viscosity, 1-2 U of DNase 1 (Thermo Scientific, Rockford, IL) was added to
break down nucleic acids. The lysate was further incubated for an additional 30-60 min under the
previous conditions. To analyze the total protein content in the processed lysate, 500 µL aliquots
of each lysate were taken and prepared for SDS-PAGE (as described in Section 3.6.1).
3.6 SDS-PAGE
3.6.1 SDS-PAGE procedure
For samples loaded on a gel for SDS-PAGE, 50 % (v/v) of the sample is made up of 2X
sample buffer prepared as follows: 5.0 mL deionized water, 2.5 mL 0.5 M Tris-HCl (pH 6.8)
(Fisher Scientific, Fairlawn, NJ), 2.5 mL of 80 % (v/v) glycerol in water, 7 mg bromophenol blue
(Fisher Biotechnology, Fairlawn NJ), 0.4 g sodium dodecyl sulfate (SDS) (Fisher Scientific,
Fairlawn, NJ) and 0.31 g dithiothreitol (DTT) (Fisher Scientific, Fairlawn, NJ)). Depending on the
lysate sample dilution required, the remainder of the sample was made up of the lysate sample and
1X PBS. A 1-L of 1X PBS contains 8 g NaCl, 0.2 g KCl (Fisher Scientific, Fairlawn, NJ), 1.4 g
Na2HPO4 and 0.2 g KH2PO4 in 1 L of dH2O. For example, a 4X lysate dilution would contain 25
% lysate sample, 25 % PBS and 50 % sample buffer (by volume). The sample now containing the
lysate, sample buffer and PBS were placed in a water bath at ~ 90°C for at least 3 min. After
incubation in the water bath, the samples were ready to be run on a gel.
For each of the samples, 10 µL were loaded onto the wells on the gel. Depending on the
gel, 3 µL of a molecular weight ladder (Precision Plus Protein Dual Colour Standards, Bio-Rad)
was added to one lane on the gel while a series of standards containing between 150 ng to 1000 ng
45
of Bovine Serum Albumin (BSA) (Thermo Scientific, Rockford, IL) were also loaded onto the gel.
The molecular weight ladder enables the identification of proteins run on the gel based upon their
molecular weight. Since the amount of BSA in each of the standards is known, a standard curve
can be obtained for each gel and used to quantify other proteins via densitometry measurement
using ImageJ software [94].
After each sample was loaded onto the gels, the gel holder and tank were filled with 1X
SDS-PAGE Running Buffer. 1X SDS-PAGE Running Buffer contains 3.03 g of Tris base, 14.4 g
of glycine (Fisher Scientific, Fairlawn, NJ), 1 g of SDS, a pH of 8.3 (adjusted using HCl) in a final
volume of 1 L (completed to final volume using deionized water). The gels were then run for up
to 1 h using a voltage that ranged from 120-180 V (Bio-Rad Mini-Protein Tetra Cell, 4-Gel
System).
Following the protein migration, the gels were placed in microwave-safe containers filled
with 100 mL of tap water and microwaved for 55 s at 900 W (Salton). The washing step was
repeated two more times for a total of three washes in order to remove SDS from the gels. After
the third wash, water was discarded and the gel was stained with 10 mL of PAGE-Blue Protein
Staining Solution (Thermo Scientific, Rockford, IL) and microwaved for 7 s three times while
ensuring that the solution was mixed effectively in between each successive microwave run. The
gel was then placed on a rocking platform (Lab Rotator Model: 2304, Thermo Scientific) for 20-
30 min. The PAGE-Blue solution was then recovered and the gel was re-suspended in deionized
water and placed on a rocking platform for 15 min. After 15 min, the water was discarded and
replaced with fresh water. The gel was left overnight and then scanned (Gel Doc Imager system,
Bio-Rad) using the Coomassie Blue setting.
46
3.6.2 SDS-PAGE analysis
In order to determine the extent of lysis that had occurred during an experiment, an
assumption was made that the amount of fractional lysis that had occurred is directly proportional
to the fractional amount of protein that had been released into the soluble phase from the E. coli
cells. For this assupmtion to hold true it was assumed that there was no degredation of proteins
occuring during the experiment and that lysis of cells results in protein release.
The fractional lysis observed in each experiment was determined by taking the pixel
density of all proteins in the soluble phase (e.g. lane S1 of the gel in Figure 3-1) and the total phase
(e.g. lane T1 of the gel in Figure 3-1) separately and then dividing the two quantities as shown in
Equation (3.1)). ImageJ software enables the user to quantify the amount of protein that has been
run on a gel through densitometry. ImageJ can be accessed and downloaded at
http://imagej.nih.gov/ij/download.html. Please note that Equation (3.1) is used for all ImageJ
analysis throughout this thesis.
Fractional lysis =
∑ Pixel Density of proteins in soluble lane "S"
∑ Pixel Density of proteins in total lane "T"
(3.1)
Figure 3-1 provides an illustrative example of the calculation performed using Equation
(3.1). In order to calculate the fractional lysis for sample 1, the sum of the pixel densities of the
bands in S1 (soluble protein portion of sample 1 illustrated by the dashed box) is divided by the
sum of the pixel densites of the bands in T1 (total protein of sample 1 illustrated by the solid box).
In order to determine the fractional lysis for sample 2 the sum of the pixel densities for S2 and T2
would be used in the above calculation instead.
47
Figure 3-1: A typical SDS-PAGE of E. coli proteins. L denotes “protein ladder”, SD denotes a BSA standard, PP
denotes the purified protein of interest, T denotes total protein fraction, S denotes soluble protein fraction, P denotes
insoluble protein fraction. The proteins enclosed within the solid box represent the “total” proteins for sample 1 while
the proteins enclosed within the dashed box represent soluble proteins for sample 1. Proteins indicated by the arrow
are removed from the quantification analysis. Numerical values denote the cell lysate sample.
It should also be noted that not all of the proteins displayed on each of the gels were
included in the analysis for determining the amount of fractional lysis. Figure 3-1 provides an
illustrative example of such a case. For all gels that were analysed, it was observed that a set of
proteins of c.a. 23-24 kDa (indicated by the arrow in Figure 3-1) did not shift from the total phase
to the soluble phase at a rate comparable to that of the other proteins as cells underwent lysis. It is
hypothesised that these proteins may be membrane-associated and that they remain attached to the
membrane of the E. coli cell after lysis and therefore would only be seen in the insoluble phase
after centrifugation. Because of this, these proteins cannot accurately represent the extent at which
lysis occurs. For this reason, they have been omitted from the calculation of fractional lysis for all
gels shown in future sections in order to maintain consistency.
3.7 Western blot analysis
Western blots were performed over the course of experimentation to determine the
presence of a specific protein of interest [95]. The protein samples are treated and processed as
48
they would be for SDS-PAGE. The migrated proteins on the gel are then transferred to a membrane
where an antibody is added which binds to an epitope on the protein of interest. The antibody is
conjugated to an enzyme that is able to be visualized with the addition of a substrate.
The first step in the process was to prepare and run protein samples on a gel as described
in Section 3.6.1. After protein migration, the gel was washed for 15 min in deionized water. The
proteins were then transferred to polyvinylidene fluoride (PVDF) membranes (Pall Corporation,
East Hills, NY) under 350 mA for 60 min while submerged in a Transfer buffer prepared by mixing
100 mL methanol (VWR BDH Chemicals, Westchester PA), 100 mL of 100 mM CAPS (Sigma
Aldrich, Oakville, ON) solution and 800 mL of deionized water per litre of buffer solution.
After transferring the proteins from the gel to the membrane, the membrane was put into a
container and washed with TBS-T on a rocking platform for 10 min. TBS-T contains 100 mL of
TBS 10X (6.05 g Tris Base, 8.76 g NaCl in 1 L of H2O, pH adjusted to 7.5 with 1 M HCl), 900
mL of deionized water and 1 mL of 50 % aqueous Tween 20 (Fisher Scientific, Fairlawn, NJ) for
each litre of solution. After 10 min, the TBS-T was drained and replaced with blocking solution.
The blocking solution contains 100 mL of TBS-T 1X and 1 g of BSA (Fisher Scientific, Fairlawn,
NJ) per 100 mL of blocking solution. The membrane was then left to incubate in the blocking
solution at room temperature for at least 45 min on a rocking platform.
The blocking solution was removed and the membrane was washed three times for 5 min
with TBS-T on a rocking platform. After the third wash, buffer was drained and 2 µL of 1 % Anti-
HIS horseradish peroxidase (HRP) antibody (Bethel Rabbit anti-6xHIS-HRP Conjugated,
Cedarlane) and 20 mL of blocking buffer were added to the membrane which was then placed on
a rocking platform. After 45 min of incubation, the antibody solution was recovered and the
membrane was washed three times for 5 min with TBS-T while on a rocking platform.
49
Next, the membrane was exposed to Western Blotting Substrate which was prepared using
a Bio-Rad Conjugate Substrate Kit (#170-6431). For each membrane, the Western Blotting
Substrate was composed of 500 µL of HRP CDB (Bio-Rad), 4500 µL of deionized water, 30 µL
of HRP color reagent B (Bio-Rad), and 1 mL of HRP color reagent A (Bio-Rad). The substrate
was allowed to react with the membrane for a maximum of 45 min until bands formed. The
membrane was then washed 1-2 times with deionized water in a container on a rocking platform
for 10 min to stop the colorimetric reaction. The membrane was scanned using a Bio-Rad Gel Doc
system.
It should be noted that western blots were used in order to detect the presence of the protein
of interest in each fermentation. Figure 3-2 provides an example of an SDS-PAGE image and a
corresponding western blot (performed for the protein of interest) of a lysis experiment. As can be
seen in Figure 3-2, a western blot is required to ensure that there is indeed the protein of interest
in this run since the protein of interest cannot be seen with certainty in pane A of Figure 3-2.
Figure 3-2: An example of results obtained for SDS-PAGE (A) and western blot (B) for a lysis experiment. In pane
A, the arrow indicates the location of the protein of interest. Pane B illustrates a western blot performed for the protein
of interest.
Mw
(kDa)
250
150
100
75
50
37
25
20
15
10
B
A
50
The amount of protein of interest released is considered to be proportional to the amount
of total protein released. This can be verified by comparing the relative intensities of the western
blot in pane B of Figure 3-2 to the intensities of proteins found in the SDS-PAGE of pane A of
Figure 3-2. It should be emphasised that identical samples were loaded for both figures. It can be
seen that the pattern of intensities from lane to lane is identical between the proteins found in pane
A and the protein of interest illustrated in pane B. This identical pattern indicates the
proportionality of intensity between the protein of interest and the amount of protein within each
sample. Therefore, the amount of release of the protein of interest can be inferred based upon the
total release of proteins. Western blots were not used to quantify the amount of fractional lysis
which occurred during experimentation since they would be too costly and time consuming to
produce for each run. Therefore, the quantification of total fractional lysis was performed using
the results obtained from SDS-PAGE. Consequently, western blots are not included in future
sections.
51
CHAPTER 4: EVALUATION OF LYSIS
TECHNIQUES
The following chapter describes the experiments and analysis performed to determine the
optimal lysis technique for the release of proteins from E. coli cells. The lysis methods that were
examined in this investigation include: 1) high pressure homogenization (Section 4.1), 2) bead
milling (Section 4.2), 3) sonication (Section 4.3), and 4) synergistic lysis using lysozymes and
Triton X-100 (Section 4.4).
4.1 High pressure homogenization
One method, used industrially, for cell lysis is high pressure homogenization. This lysis
method can process up to 50 000 L/h of cell suspension at 7 MPa (21 000 L/h at 25 MPa) (based
on the Gaulin 185Q with poppet type valves from APV Homogenizers [99] ). In this section, the
ability and efficiency of a high pressure homogenizer for E. coli cell lysis were studied. High
pressure homogenization, also known as French pressure cell press or simply French press, is
considered to be an excellent method for cellular disruption [39]. High pressure homogenizers are
able to disrupt E. coli cells by forcing the cells through a very small orifice via a very high
differential pressure. The cell lysates are thereby exposed to an extremely high shear rate and, as
a result, cell disruption and lysis occur. Since E. coli cells are classified as gram-negative bacteria,
they are easier to lyse through high pressure homogenization compared to gram-positive bacteria
due to their thicker peptidoglycan layer [55].
In this investigation, experiments were performed with an Emulsiflex C5 (Avestin Inc.,
Ottawa, Canada) high pressure homogenizer to determine the amount of lysis that could be
achieved with equipment that is considered to be the industry benchmark for lysis efficiency. This
52
allowed for a lysis baseline to be set and used as a comparison between the different lysis
techniques for E. coli cells. As cell lysis is proportional with the number of passes through the high
pressure homogenizer [57], a series of experiments were performed to determine the extent of lysis
as well as the potential negative shearing effect on released proteins resulting from several passes
through the apparatus.
4.1.1 Results and discussion
Each sample of the lysis experiments was analysed to determine their protein content using
SDS-PAGE. A typical gel obtained from SDS-PAGE analysis is shown on Figure 4-1. The “total”,
“soluble” and “pellet” samples for each successive pass through the homogenizer were run on the
gel.
Figure 4-1: SDS-PAGE of E. coli proteins. E. coli cells were grown in a 1-L bioreactor (750 mL working volume)
with a pre-induction and post-induction temperature of 28°C. A 50 mL cell suspension was lysed using an Avestin
Emulsiflex C5 high pressure homogenizer for 3 passes at 7 mL/min. L denotes “protein ladder”, SD denotes a BSA
standard, T denotes total protein, S denotes soluble protein, and P denotes insoluble protein. Numerical values indicate
the number of passes through the high pressure homogenizer for that particular sample. Pane A represents a 10 %
(w/v) E. coli run. Pane B represents a 20 % (w/v) E. coli run. Both runs were performed at a 10X dilution on each gel.
Please note that the antibody of interest is not displayed in these gels due to the fact that the runs completed above
were only performed to screen for a visualization of the total amount of E. coli lysis.
Figure 4-1 shows that after a single pass for both the 10 % and 20 % (w/v) samples, almost
100 % of the available proteins have been released from the cells. This is inferred since the lanes
containing the soluble proteins, S1, from both samples after the first pass are almost identical in
53
intensity when compared to the lane that contains the total amount of proteins, T1. It can also be
noted that the intensity of the insoluble phase (pellet in lane P1) decreases substantially after one
pass indicating that lysis is occurring; the proteins are shifting from the insoluble phase (proteins
trapped within the cell) to the soluble phase (proteins released into the extracellular environment).
The amount of proteins depicted in Figure 4-1 was quantified and analyzed using ImageJ analysis
as described in Section 3.6.1. The results of the analysis can be seen in Figure 4-2. It should be
noted that the fractional lysis of the 10 % (w/v) sample after 2 passes was greater than 1 which is
not possible. This situation may have occurred due to the inherent error of the quantification
method compounded by the fact that the intensities of the two quantified lanes (S2 and T2 in pane
A of Figure 4-1) are visually almost identical.
Figure 4-2: Fractional release of proteins from E. coli cells based on the number of passes through an Avestin
Emulsiflex C5 high pressure homogenizer at 7 mL/min. One run for each cellular concentration (10 % (w/v) and 20
% (w/v)) was performed and compared.
In literature, it has been shown that an increase in the number of passes of an E. coli
suspension through a high pressure homogenizer leads to an increase in the amount of lysis and
therefore an increase in the amount of proteins released [57]. However, it can be seen from Figure
0
0.2
0.4
0.6
0.8
1
1.2
0 1 2 3
Fra
ctio
nal
Lysi
s
Number of Passes
10% 20%
54
4-2 that total cell lysis seems to occur and that nearly all of the proteins are released from the cells
during the first pass through the high pressure homogenizer. Based on this observation, higher
density cell suspensions, higher flow rates and lower pressures could be tested to observe if a high
amount of lysis would still be maintained. This would allow for the lysis of a greater amount of E.
coli cells over a shorter period of time at a reduced operating cost.
4.1.2 Summary
It can be concluded that for cell suspensions at 10 % and 20 % (w/v), nearly 100 % of the
proteins that are located within the E. coli cells are released after one pass through the Avestin
Emulsiflex C5 high pressure homogenizer at a flow rate of 7 mL/min and pressure of 152 MPa.
Therefore, subsequent passes were not needed to obtain a maximum amount of soluble proteins.
Unfortunately, due to the lack of availability of the high pressure homogenizer, only one set of
experiments was performed. Other variables that could potentially be tested in the future would be
a wider range of cell concentrations of the E. coli cell suspension, the flowrate, or homogenizers
that contain a differently shaped impact ring. Furthermore, to reduce processing time of larger
sample volume, higher capacity units can be used. The yield of fractional lysis could also be
improved by pre-treating the E. coli cells either chemically or enzymatically in order to weaken or
compromise the cellular structure prior to high pressure homogenization [61].
4.2 Bead milling
The following section investigates E. coli lysis using bead milling. During bead milling, a
cell suspension is mixed at high speeds with glass beads. With the vivid agitation, the glass beads
collide with the cells causing lysis and the release of proteins. For the series of bead milling
experiments, a 90-mL BioSpec BeadBeater was used to test: 1) the extent of lysis using a
laboratory batch-scale bead milling, 2) the effect of lysing cells sampled during the exponential
55
and stationary growth phases, 3) the required lysis duration, and 4) the effect of the size of the
glass beads.
4.2.1 Bead milling: effect of bead to lysate ratio
It has been determined in literature that the bead to lysate ratio of a bead milling experiment
has a significant impact on the amount of cellular lysis that can occur. It has been found that in
order to achieve a maximum amount of cellular lysis, 70 % of the entire chamber volume must be
occupied by glass beads [69]. However, during preliminary testing with a 90-mL BioSpec
BeadBeater with 70 % glass bead loading, the apparatus was not able to impart sufficient velocity
to the glass beads to cause cell lysis. Therefore, the bead loading percentage was lowered to 50 %.
Three reasons motivated lowering the maximum bead loading to 50 %: 1) the apparatus was able
to function properly at this ratio, 2) samples were able to be taken directly from the unit during
time trials (at 60 % loading, sampling proved to be very difficult), and 3) 50 % bead loading is
recommended by the manufacturer to achieve an advertised cell lysis of 90 % [100]. Therefore,
results presented in this section were obtained with a glass bead loading percentage of 50 % which
corresponds to a cell suspension:glass bead volume ratio of 1:1.
4.2.2 Bead milling: cell growth cycle point and time
The main objective of this section was to investigate if the lysis efficiency was affected by the
growth stage of the culture, i.e. the moment at which cells were harvested. Lysis was performed
using bead milling for fermentation broth samples collected at different time points in cellular
fermentations. The time points chosen were representative of the main two phases in the
fermentation: the exponential phase and stationary phase. It should be noted that all samples were
taken during stationary and exponential phases that occurred after the induction of the protein of
interest. These experiments were performed to see if cells in an exponential growth phase would
56
be more susceptible to lysis through bead milling compared to cells in a stationary phase. Higher
susceptibility in the exponential phase has been demonstrated during experiments with high
pressure homogenization [39], [101], osmotic shock [39], [102], enzymatic lysis [39], [103], [104],
ultrasound and bead milling [39], [105]. E. coli cells in exponential phase are continuously
dividing, which make their peptidoglycan structure weaker compared to cells in a stationary phase
[39]. Therefore, cells in the exponential phase should lyse more easily [39]. It was then desired to
determine if there would be any advantage of lysing the E. coli cells at a particular point in the
fermentation.
The final objective of this section was to determine the effect that time had on the extent
of E. coli lysis during bead milling. The purpose of studying this parameter is to determine the
optimal amount of time required to efficiently lyse a sample. It has been demonstrated in literature
that an increase in lysis time during bead milling results in an increase in the amount of lysis
observed [69], [74]. The amount of time required for lysis can have a direct impact on the required
volume and the flow rates when designing and operating a continuous bead milling apparatus.
4.2.2.1 Results and discussion
Growth curves from Fermentations 1 to 3 and sample points harvested for lysis are illustrated
in Figures 4-3 to 4-5. From these figures, each of the samples taken was either classified as being
harvested from the exponential or stationary phase as described in Table 4-1.
57
Figure 4-3: E. coli TG1 growth curve from batch culture in modified TB medium (4 g/L glycerol at 28°C) with respect
to fermentation time. The batch phase, with a volume of 750 mL, was performed for 17 h prior to induction with
IPTG. Cell samples for lysis were taken at 4, 12 and 24 h post induction as indicated by dashed lines.
Figure 4-4: E. coli TG1 growth curve from fed-batch culture in modified TB medium. The batch phase, with a volume
of 400 mL, was performed in TB media, 4 g/L of glycerol at 28°C for 20.5 h prior to induction with IPTG. Fed-batch
medium (150 g/L glucose, 15 g/L glycerol) was initiated 3.5 h prior to induction with 300 mL total medium added
continuously over 3 h. Cell samples for lysis were taken at 4, 8 and 20.5 h after induction as indicated by dashed lines.
0
10
20
30
40
50
60
70
0 5 10 15 20 25 30 35 40 45
OD
600
Fermentation Time (h)
Induction
0
10
20
30
40
50
60
70
0 5 10 15 20 25 30 35 40 45
OD
600
Fermentation Time (h)
Induction
58
Figure 4-5: E. coli TG1 growth curve from fed-batch culture in modified TB medium. The batch phase, with a volume
of 400 mL, was performed in TB media, 4 g/L of glycerol at 37°C for 16 h prior to induction with IPTG. Fed-batch
medium (150 g/L glucose, 15 g/L glycerol) was initiated 3.5 h prior to induction with 300 mL total medium added
continuously over 3 h. After induction, the temperature was decreased to 28°C. Cell samples were taken at 4, 8 and
20.5 h after induction as indicated by dashed lines.
Table 4-1: Cellular growth phase classification for cell fermentations 1, 2 and 3. Classification is based upon the data
collected and displayed in Figures 4-3 to 4-5. Samples taken during the increasing linear portion of each figure were
classified as being in the exponential phase. Samples taken during the flat portion of each figure were classified as
stationary.
Fermentation Number Sample Time after Induction (h) Phase Classification
1 4 Stationary
1 12 Stationary
1 24 Stationary
2 4 Exponential
2 8 Exponential
2 20.5 Stationary
3 4 Exponential
3 8 Exponential
3 20.5 Stationary
0
10
20
30
40
50
60
70
80
90
0 5 10 15 20 25 30 35 40
OD
600
Fermentation Time (h)
Induction
59
Figure 4-6 provides a comparison of SDS-PAGE gels for three samples taken at different
times after induction. “Total”, “soluble” and “pellet” samples were run on the gel which provide
the total, soluble and insoluble fractions of each sample. Comparisons were then made between
the “total” and “soluble” samples in order to prepare Figure 4-7.
Figure 4-6: A-C: SDS-PAGE of E. coli proteins from Fermentation 3. E. coli cells were grown in a 1-L bioreactor
(750 mL working volume) with a pre-induction and post-induction temperature of 28°C. Cells were lysed using a 90-
mL BioSpec BeadBeater chamber, 50 % cell suspension, 50 % 0.1 mm beads, between 0 and 9 min. L denotes “protein
ladder”, SD denotes a BSA standard, PP denotes the purified protein of interest, T denotes total protein, S denotes
soluble protein, and P denotes insoluble protein. Numerical values indicate at what time of bead milling the sample
was taken. A: Cells were harvested 4 h after induction. B: Cells were harvested 8 h after induction C: Cells were
harvested 20.5 h after induction.
Figure 4-7 provides a comparison of the fractional cellular lysis achieved with respect to
the growth phase and milling time of E. coli cells in a 90-mL BioSpec BeadBeater chamber as
described in Section 3.5.2. It should be noted that the data displayed in Figure 4-7 has been pooled
from bead milling experiments performed on E. coli cells from Fermentations 1 to 3. The results
of each lysis experiment were tabulated in Figure 4-7 based on the classification of each sample
as described in Table 4-1.
60
Figure 4-7: A comparison of the amount of fractional lysis of E. coli cells over 9 min of bead milling. Cells were
harvested during either exponential or stationary phases of the fermentation. E. coli cells were lysed using a 90-mL
BioSpec BeadBeater chamber and sampled at 3, 6 and 9 min. Cellular lysis was performed using 0.1 mm beads (50
% glass beads: 50 % cellular lysate). Samples were run on SDS-PAGE and evaluated through densitometry using
ImageJ software to determine the fractional amount of lysis performed. Fractional conversion of lysis was determined
by dividing the amount of protein present in the “soluble” lanes in each gel by the amount of protein present in each
corresponding “total” lane. Error bars indicate 1 standard deviation based on either 4 or 5 trials. Each exponential
data set contained 4 data points while each stationary data set contained 5 data points.
It has been seen in literature that E. coli cells harvested during the exponential phase are
more susceptible to lysis compared to E. coli cells harvested during the stationary phase [39],
[105]. Figure 4-7 demonstrates the amount of fractional lysis obtained during the above
experiment: 0.79 0.12 after 3 min, 0.648 0.28 after 6 min and 0.54 0.08 after 9 min for cells
harvested in the exponential phase and 0.67 0.05 after 3 min, 0.65 0.08 after 6 min and 0.56
0.08 after 9 min for cells harvested in the stationary phase. Using Quantum XL (Accessed through
www.sigmazone.com/QuantumXL.htm), it has been determined that there is no significant
difference between lysing the cells in the exponential phase or in the stationary phase since the p-
value for this variable is greater than 0.1 (0.585). Since it has been found that there is no significant
difference between lysing cells harvested from either phase it is suggested to harvest E. coli cells
from the stationary phase for lysis applications. Cells that are harvested during the stationary phase
0
0.2
0.4
0.6
0.8
1
0 3 6 9
Fra
ctio
nal
Co
nver
sio
n
Lysis Time (min.)
Exponential Stationary
61
have completed cellular growth and have therefore produced the maximum concentration of
protein possible (in the case of a growth associated proteins) [11]. Furthermore, the fermentation
would not have to be as closely monitored as harvesting cells from the exponential phase where a
point representing the maximum protein concentration during the exponential phase would have
to be identified [11].
In addition, it can be observed from Figure 4-7 that the average amount of lysis found in
samples taken after 3 min show an apparent decrease in the overall lysis. Using Quantum XL
software, this result has been determined to be statistically significant since the p-value for time
for exponential and stationary samples harvested between 3 and 9 min was determined to be less
than 0.05 (0.005). This contradicts what is found in literature that an increase in the bead milling
time results in an increase of cellular lysis [69], [74]. Since the extent of lysis is calculated based
on the quantification of proteins, a decrease in proteins would cause this apparent decrease in lysis.
As bead milling time increases, proteins may be destroyed due to the repeated collisions, and shear
caused by the glass beads. High temperatures could also contribute to protein degradation.
Therefore, the temperature of the bead suspensions for all runs were recorded and illustrated in
Figure 4-8.
62
Figure 4-8: Effect of bead milling time on cell lysate temperature. The data was compiled by measuring the
temperature of 9 runs of bead milling lysates, using 0.1 mm glass beads (BioSpec) over 0, 3, 6 and 9 min.
From Figure 4-8, it can be seen that the highest average temperature recorded was 29°C
which was recorded after a milling time of 6 min. Since this temperature is relatively low, it is
highly unlikely that the proteins were degraded due to the temperature. However, it can be noted
that a significant amount of lysis does occur after the first 3 min of milling, (0.79 0.12 and 0.67
0.05 for exponential and stationary phases respectively), prior to the observed decrease in lysis.
This is an indication that cells can be still be lysed very quickly and efficiently using this technique
as long as the lysis is performed for no longer than 3 min.
4.2.2.2 Summary
It was determined that there was no significant difference in lysing cells that have been
harvested in either the exponential or stationary phase with the obtained data: 0.79 0.12 after 3
min, 0.648 0.28 after 6 min and 0.54 0.08 after 9 min for cells harvested in the exponential
phase and 0.67 0.05 after 3 min 0.65 0.08 after 6 min and 0.56 0.08 after 9 min for cells
harvested in the stationary phase. Therefore, it is recommended to lyse cells that have been
0
5
10
15
20
25
30
35
0 3 6 9
Tem
per
ature
(°C
)
Milling Time (min)
63
harvested in the stationary phase in order to ensure the highest concentration of the protein of
interest as possible.
Finally, an apparent decrease was seen in the amount of fractional lysis when the duration
of lysis is extended past 3 min. This could be an indication that some of the released proteins are
degrading during the bead milling process. It should be noted that the amount of fractional lysis
achieved after 3 min of milling is considerable (0.79 0.12 and 0.67 0.05 for the exponential
and stationary phases respectively), indicating that bead milling can be a viable lysis technique.
Currently it is recommended to mill for no longer than 3 min, however, further testing between 0
and 3 min of bead milling should be performed to determine if the optimal lysis treatment time
could be lowered to less than 3 min.
4.2.3 Bead milling: effect of bead size
The objective of this section was to investigate the effect of different glass bead diameters
on the lysis of E. coli cells using bead milling. It has been recommended by the manufacturer
(BioSpec) that the optimal size of glass beads is 0.1 mm diameter to lyse bacterial cells [100].
However, it has been demonstrated in literature that a glass bead size of 0.75 mm instead provides
an optimal amount of lysis [69]. Therefore, a comparison should be made between the two bead
sizes in order to determine the true optimal size. However, for this experimentation it was not
possible to obtain 0.75 mm diameter glass beads due to availability. Instead, three types of beads
were used: 0.1 mm and 0.5 mm glass beads from BioSpec which could be purchased for $79.2/kg,
and a mixture of lower-cost glass beads with an average size of 0.3 mm from McMaster-Carr
(Elmhurst, Illinois) which could be purchased for $9.50/kg. It is to be noted that at the time of the
experiment, the company wanted to have tests performed on bacteria producing 14-2M protein.
Previous results shown in Sections 4.1.1 and 4.2.1 were based on 2-3P proteins.
64
4.2.3.1 Results and discussion
Figure 4-9 provides a visual representation of the glass beads utilized during this
experimentation. Characteristics and size distribution of the particles are presented in Figures 4-10
to 4-12.
Figure 4-9: A-C Microscopic images of glass beads used for the operation of milling E. coli. Images were taken using
a light microscope. A: 0.1 mm diameter glass beads (BioSpec) B: 0.3 mm glass beads (average size) (McMaster-Carr)
C: 0.5 mm glass beads (BioSpec) Each division on each scale represents 0.1 mm. Please note, there is no void in the
centre of each glass bead; the void-like appearance in the centre of each bead is due to the reflection of light. Each
image has a magnification of 20X.
Figure 4-10: Particle size distribution of the “0.1 mm” diameter beads (BioSpec). Beads were characterized by
measuring 250 beads and calculating the area of the beads using the software ImageJ. The area of each particle was
then converted to a respective diameter. For the “0.1 mm” the average particle size was determined to be 0.097 mm
with a standard deviation of 0.041.
65
Figure 4-11: Particle size distribution of the “0.3 mm” diameter beads (McMaster-Carr). Beads were characterized by
measuring 250 beads and calculating the area of the beads using the software ImageJ. The area of each particle was
then converted to a respective diameter. For the “0.3 mm” the average particle size was determined to be 0.318 mm
with a standard deviation of 0.061.
Figure 4-12: Particle size distribution of the “0.5 mm” diameter beads (BioSpec). Beads were characterized by
measuring 250 beads and calculating the area of the beads using the software ImageJ. The area of each particle was
then converted to a respective diameter. For the “0.5 mm” the average particle size was determined to be 0.539 mm
with a standard deviation of 0.037 mm.
It can be observed from Figures 4-10 and 4-12 that the 0.1 mm and 0.5 mm glass beads
have a narrow distribution with sizes ranging from 0.07-0.14 mm and 0.49-0.58 mm for the 0.1
66
mm and 0.5 mm glass beads respectively. The distribution 0.3 mm glass bead mixture in Figure
4-11 has a larger distribution: 0.25-0.42 mm.
E. coli cells were lysed over a span of 6 min using glass beads of 0.1 mm, 0.3 mm or 0.5
mm in diameter. Samples taken at 0, 3 and 6 min were run on SDS-PAGE in order to visualize the
amount of overall proteins (total) and the amount of corresponding soluble and insoluble proteins.
Typical gels can be seen in Figure 4-13 and were analyzed using ImageJ densitometry (as
described in Section 3.6.1) in order to produce the results seen in Figure 4-14.
Figure 4-13: A-C: SDS-PAGE of E. coli proteins. E. coli cells were grown in a 7.5-L bioreactor (working volume 5
L) bioreactor with a pre-induction and post-induction temperatures of 32°C and 28°C respectively. Cells were lysed
using a 90-mL BioSpec BeadBeater, 50 % cell suspension, 50 glass beads, between 0 and 6 min. PP denotes the
purified protein of interest, T denotes total protein, S denotes soluble protein, and P denotes insoluble protein.
Numerical values indicate the time the samples were taken. A: Cells were lysed using 0.1 mm glass beads B: Cells
were lysed using 0.3 mm (average size) glass beads C: Cells were lysed using 0.5 mm glass beads.
67
Figure 4-14: Comparison of the amount of fractional lysis of E. coli cells over 6 min of bead milling. E. coli cells were
lysed using a 90-mL BioSpec BeadBeater and sampled at 0, 3 and 6 min. Cellular lysis was performed using 0.1 mm,
0.3 mm (average bead size), or 0.5 mm glass beads (50 % glass beads, 50 % cellular lysate). Samples were run on
SDS-PAGE and evaluated through densitometry using ImageJ software to determine the fractional amount of lysis
performed. Fractional conversion of lysis was determined by dividing the amount of protein present in the “soluble”
lanes in each gel by the amount of protein present in each corresponding “total” lane. Error bars indicate 1 standard
deviation based on 3 trials.
Results of Figure 4-14 show that there is no significant difference between lysing cells
using 0.1 mm beads, 0.3 mm (average bead size) beads or the 0.5 mm beads. The amount of
fractional lysis obtained after 3 min of lysing was, on average, 0.63 0.08, 0.70 0.10 and 0.55
0.23 for the 0.1 mm, 0.3 mm and 0.5 mm beads, respectively. This implies that beads larger than
0.1 mm and smaller than 0.5 mm can be used in order to lyse E. coli cells. Larger beads can be
advantageous because they are easier to handle and they have a lower probability of being trapped
under seals and damaging equipment. Furthermore, results in Figure 4-14 show that the 0.3 mm
beads were able to lyse E. coli cells either as well or better than 0.1 mm beads or 0.5 mm beads.
Since the 0.3 mm beads are a mixture of multiple bead sizes they can be purchased at a fraction of
the cost of beads having a narrow size distribution (ex. 0.1 mm beads or 0.5 mm beads from this
experiment). Therefore, it is recommended to use the 0.3 mm bead mixture when scaling up to
larger continuous bead milling operations in order to save on production costs.
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 3 6
Fra
ctio
nal
Co
nver
sio
n
Amount of Lysis Time (min)
0.1 mm 0.3 mm 0.5 mm
68
It can also be noted that the amount of lysis seems to plateau. This is an indication that 3
min of lysis time could be sufficient to achieve the highest quantity of lysis possible through bead
milling with these sizes. This phenomenon was also described in Section 4.2.2 with bead milling
experiments pertaining to lysis with 0.1 mm glass beads.
It has been reported in literature that using 0.75 mm glass beads leads to a greater amount
of lysis (80 %) compared to glass beads ranging from 0.25 mm to 0.5 mm (65 % and 70 %
respectively) [69]. It should also be noted that it was found in literature that bead sizes greater or
equal to 1 mm in diameter resulted in E. coli lysis of less than 50 %. It is difficult to compare these
results to the ones stated above since different milling units and different operating conditions were
used. However, if it were possible a 0.75 mm bead size should be investigated in future
experiments.
It is important to note that the amount of lysis occurring at time 0 min cannot be assumed
to be 0. Indeed, Figure 4-14 illustrates that the average amount of fractional cellular lysis of E. coli
at time 0 was 0.4. This indicates that there is another factor contributing to lysis prior to bead
milling. One contributing factor to this pre-lysis may be the freeze-thaw process. When cells are
frozen, ice crystals can form which could potentially cause some cells to lyse. This phenomenon
is prevalent when freezing cells at higher temperatures such as -20°C compared to lower
temperatures such as -80°C. At -20°C, the freezing process is very slow [79] which allows for a
higher amount of ice crystals to form within the cells promoting a larger amount of lysis. At lower
temperatures such as -80°C, the cells freeze very quickly, which reduces the formation of ice
crystals within the cells. Since the cells were stored at -20°C, there is a possibility that the above
freezing phenomena could contribute to the lysis seen prior to bead milling as indicated in Figure
4-14 [79].
69
Another potential explanation for the lysis prior to bead milling could be due to the addition
of EDTA. The addition of EDTA was used in this experimentation to prevent protein degradation
[106] from proteases. However, the addition of EDTA can also destabilize the outer membrane of
gram-negative bacteria like E. coli which can lead to cellular lysis. This form of lysis could also
possibly contribute to the phenomena demonstrated in Figure 4-14 [41], [42].
4.2.3.2 Summary
In conclusion, results have shown that the bead diameter has an insignificant effect on the
lysis of cells such that the bead mixture having an average diameter of 0.3 mm can be used for the
efficient lysis of E. coli cells. Since the 0.3-mm diameter beads are substantially less expensive
than the two more uniform sizes of glass bead ($9.50/kg compared to $79.2/kg), it is recommended
to use the 0.3 mm diameter beads when scaling up the bead milling operation. Results also showed
that the optimal amount of lysis was achieved after 3 min (0.63 0.08, 0.70 0.10 and 0.55 0.23
for the 0.1 mm, 0.3 mm and 0.5 mm beads, respectively). This result was also seen earlier in the
lysis experimentation over time with bead milling at different cellular growth phases. It was also
seen that the amount of lysis plateaus for each of the respective bead sizes after 3 min of bead
milling. It is recommended to investigate the lysis of E. coli cells using 0.75 mm beads since this
was the optimal bead size determined in literature [69]. Finally, it was noted that there were non-
negligible amounts of lysis prior to bead milling at 0 min. This amount of pre-lysis could be
attributed to thawing the cells or the addition of EDTA.
4.3 Sonication
A series of experiments were performed to elucidate the efficiency of sonication on the
lysis of E. coli cells. Sonication, using a sonic probe to generate sound energy usually within a
range of 20-50 kHz, is able to disrupt the structure of cells through the formation of violent
70
implosions of small bubbles and cavitation within the sample. The energy of the sonic waves is
able to cause a disruption of the intramolecular forces that provide the integrity for the cellular
wall. Four parameters have been tested in order to determine their potential effect on the cellular
lysis via sonication: output control (sonicator power input), duty cycle (% of time the sonicator is
on during the experiment), optical density (OD) and lysis time period.
4.3.1 Results and discussion
Results for the fractional cell lysis for each run of the experimental design are presented in
Table 4-2.The highest amount of fractional lysis was achieved for Run 13 with a value of 0.628.
The settings for each input parameter for Run 13 were as follows: Output Control: 6 (+1 input),
Duty Cycle: 100 (+1 input), OD: 20 (-1 input) and Time: 1 min (-1 input). In order to determine
how each of the individual parameters contributes to this finding, the experimental design in Table
4-2 was used to determine the interaction effects for each of the input parameters and to potentially
find a model to represent the lysis conversion as a function of these input parameters.
71
Table 4-2: A breakdown of the values of each parameter during the sonication experimentation: A represents Output
Control, B represents Duty Cycle, C represents OD and D represents Time (min). Y-exp. represents the average
amount of fractional lysis over 3 runs (* indicates only 2 runs used). Y-pred. represents the predicted amount of
fractional lysis using Equations (4.1) and (4.2) respectively. ^ indicates trials that were run as centre-points for the
experimental design that were unable to be reported due to gel sample loading problem.
Run A B C D y-exp. y-pred. Eqn. 1 y-pred. Eqn. 2
1 2 20 20 1 0.126 0.145 0.263
2 2 20 20 5 0.231 0.250 0.178
3 2 20 60 1 0.114 0.133 0.134
4 2 20 60 5 0.115 0.134 0.049
5 2 100 20 1 0.172* 0.153 0.165
6 2 100 20 5 0.509 0.490 0.429
7 2 100 60 1 0.062 0.043 0.036
8 2 100 60 5 0.229 0.210 0.301
9 6 20 20 1 0.228 0.209 0.280
10 6 20 20 5 0.534* 0.515 0.544
11 6 20 60 1 0.041 0.022 -0.019
12 6 20 60 5 0.282 0.263 0.245
13 6 100 20 1 0.628 0.647 0.532
14 6 100 20 5 0.408 0.427 0.446
15 6 100 60 1 0.257 0.275 0.233
16 6 100 60 5 0.028 0.047 0.148
17 4 60 40 3 ^ ^ ^
The plots of the main effects for each of the 4 input parameters: output control (A), duty
cycle (B), optical density (C) and lysis time (D) are presented in Figure 4-15.
72
Figure 4-15: The effects output control (A), duty cycle (B), optical density (C) and time (D) [min] on the fractional
amount of lysis. Parameter effects are determined through the complete model (Equation (4.1)).
An increase in the output control of the sonicator increases the power of the ultrasonic
waves used to disrupt the membrane of the E. coli cells. Results obtained in Figure 4-15 confirmed
results of previous researchers which showed that an increase in sonication power causes an
increase in the amount of cellular lysis [64], [66]–[69]. However, the effect is minor in this case
since only a 10 % increase in cellular lysis was observed when increasing the power output from
2 to 6 (80-240 W).
An increase in the duty cycle from 20 to 100 (where 20 indicates that the sonicator emits
ultrasonic waves 20 % of the time and 100 indicates 100% of the time) led to an 8 % increase in
cellular lysis. This observation makes logical sense since the sample is experiencing five times the
73
amount of sonication over the allotted amount of time allowing for a higher amount of lysis to
occur.
The optical density (OD) or cell concentration of the lysate was also varied. From Figure
4-15, it can be seen that an increase in optical density from 20 to 60 decreases the amount of
fractional lysis by about 21 %. It was noted that varying this parameter had the highest effect on
the observed amount of lysis. However, in literature it has been determined that the amount of lysis
seen during sonication is independent of the OD [39]. Therefore, this result needs to be validated
with further experimentation.
Finally, with respect to lysis time, it was demonstrated in literature that an increase in
sonication time leads to an increase in the amount of lysis obtained [58], [64], [67], [68]. As seen
in Figure 4-15, as the sonication time was increased from 1 to 5 min, the amount of lysis increased
by 9 %. It has also been reported in literature that there can be a large variation in the amount of
sonication time required to achieve maximum lysis between different batches of E. coli cells [65],
[71]. It is possible that the amount of lysis time in this experiment was not long enough to achieve
the maximum amount of lysis possible with this batch of E. coli cells.
4.3.2 Model development
The observations of the main effects seen in Section 4.3.1 were determined using Quantum
XL based on the values of the lysis experimental data for each run. The average values for each
run are demonstrated in Table 4-2. Unfortunately, the data associated with the central point could
not be used because the gel did not allow for proper quantification. Equation (4.1) predicts the
fractional lysis as a function of the 4 input parameters: output control (A), duty cycle (B), optical
density (C), and time (D). This regression equation is expressed using the coded coefficients
74
between -1 to +1 represented by the minimum and maximum of each parameter value of Table
4-2. The predicted values of the lysis fraction from Equation (4.1) for all runs are also presented
in Table 4-2.
𝑦 = 0.2477 + 0.053𝐴 + 0.0388𝐵 − 0.1068𝐶 + 0.0444𝐷 + 0.0095𝐴𝐵 − 0.0421𝐴𝐶 −0.032𝐴𝐷 − 0.0358𝐵𝐶 − 0.0374𝐵𝐷 − 0.0217𝐶𝐷 − 0.0032ABC − 0.0871ABD +0.0125ACD − 0.0006BCD + 0.0076ABCD
(4.1)
Equation (4.1) is not very useful as a predictive model since the number of fitted
coefficients is equal to the number of data points used to fit the linear model which obviously gives
a perfect prediction. Figure 4-16 presents the lysis fraction predicted by Equation (4.1) as a
function of the experimental values of the lysis fraction for the average and individual values for
each run. The lysis fraction of each run allows to visually assess the significant variability of the
experimental data.
Figure 4-16: A plot of the predicted amount of fractional lysis versus the amount of experimental lysis. Y-predicted
values are calculated from Equation (4.1). Dots indicate average y-exp. values taken as an average of 3 samples.
Diamonds indicate individual y-exp. values. Linear trend lines have been added for each set of data points. The
equation of the trend line shown is for the individual linear trend line (dotted line).
y = 0.6415x + 0.0872
R² = 0.6415
0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.800
0.000 0.200 0.400 0.600 0.800 1.000
y-p
red
icte
d
y-exp
Average
Individual
Linear (Average)
Linear
(Individual)
75
The statistical analysis allows, nevertheless, to observe which coefficient estimates from
Equation (4.1) are the most significant to estimate the fractional lysis and at the same time allows
to reduce the number of coefficients that are statistically insignificant. Figure 4-17 presents the
Pareto plot that can greatly aid in the reduction of the above equation.
Figure 4-17: Pareto plot for initial model of sonication experimentation. The bars with dots indicate coefficients where
the p-value is less than 0.05, striped histogram bars indicate coefficients where the p-value is less than 0.1 but greater
than 0.05, the checkered bars indicate coefficients where the p-value is greater than 0.1.
For the purpose of reducing the number of terms in Equation (4.1), coefficients that have a
p-value greater than 0.1 were removed from the equation since they have very little significance.
The coefficients represented by striped histogram bars (having a p-value between 0.05 and 0.1)
were left in the model. Since the duty cycle coefficient (B) had a large impact on the interaction
coefficient ABD, it was left in the equation even though it had a p-value over 0.1 (0.115). It should
be noted that there would have been great reluctance to remove the duty cycle coefficient based
upon the fact that 3 out of the 4 runs containing the greatest amount of lysis (Runs 6, 11 and 14)
are experiments that were all run at a duty cycle of 100 (the highest value for the duty cycle
0
0.02
0.04
0.06
0.08
0.1
0.12
Co
eff
Factor
76
parameter). Another regression was performed on the remaining factors and Equation (4.2) was
obtained.
𝑦 = 0.2478 + 0.0534𝐴 + 0.0385𝐵 − 0.1069𝐶 + 0.0447𝐷 − 0.0425𝐴𝐶 − 0.0874𝐴𝐵𝐷
(4.2)
Therefore, the parameters that are left in the lysis model in the case of sonication are the
output control (A), the duty cycle (B), optical density (C), the interaction between output control
and optical density (AC) and the interaction between output control, duty cycle and time (ABD).
Figure 4-18 presents the lysis fraction predicted by the reduced model as a function of the
experimental values of the lysis fraction for the average value. Fractional lysis values of individual
runs are also plotted in Figure 4-18 which allows to visually assess, once more, the significant
variability of the experimental data.
Figure 4-18: A plot of the predicted amount of fractional lysis versus the amount of experimental lysis. Y-predicted
values are calculated from Equation (4.2). Dots indicate average y-exp. values taken as an average of 3 samples.
Diamonds indicate individual y-exp. values. Linear trend lines have been added for each set of data points. The
equation of the trend line shown is for the individual linear trend line.
Given the variability of the experimental data for sonication (R2 of 0.5445), Equation (4.2)
can only be used to provide an overall tendency and to suggest further experiments in the case this
y = 0.5445x + 0.1107
R² = 0.5445
-0.100
0.000
0.100
0.200
0.300
0.400
0.500
0.600
0.700
0.000 0.200 0.400 0.600 0.800 1.000
y-p
red
icte
d
y-exp
Average
Individual
Linear (Average)
Linear (Individual)
77
method would be acceptable for lysis. The data obtained in the sonication experiments has
significant variability but some preliminary information could still be obtained on the efficiency
of this technique. The variability is due to the natural variability of the cell samples and the
sonication batch experiments. The variability can also be due to the variation in the positioning of
the tip within the sample volume. From trial-to-trial, the positioning of the tip can result in some
portions of the sample receiving higher energy intensity compared to other parts of the sample.
This variation can be due to operator error since the sonication experiments were done manually.
Also, improper mixing may have led to an improper distribution of the sonicator’s energy which
could introduce variability to the results from run to run. Finally, the gels that were used for the
protein analysis varied in staining from gel to gel, thus introducing more potential variation to the
results.
4.3.3 Summary
From the above results, it was determined that an increase in the output control from 2 to
6 (80-240 W), an increase in duty cycle from 20 to 100 % and an increase in lysis time from 1 to
5 min led to an increase in lysis of 10, 8 and 9 % respectively. The increase in lysis by increasing
each of these parameters is consistent with results reported in literature [58], [64], [66]–[69]. It has
also been found in literature that the time required to achieve maximum lysis can vary
substantially: from 20 s [65] up to 90 min [71]. One possibility for future work would be to increase
the amount of lysis time to see if this would substantially make a difference.
The variable that had the greatest effect on the amount of lysis was the optical density OD).
It was determined that increasing the OD from 20 to 60 decreased the amount of fractional lysis
by 21 %. This observation is contradictory to what has been previously determined in literature
where it has been found that changing the concentration of cells does not have an impact on the
78
amount of lysis [39], [69]. Therefore, this parameter should be further explored in order to verify
this result.
One parameter that was not tested was the effect of volume. It was seen in literature that a
decrease in volume leads to an increase in lysis [67], [70]. If the process were to be scaled-up, the
challenge of processing large volumes of cell suspension would need to be solved. The sonication
power would need to be vastly increased in order to compensate for the increase in volume.
Therefore, sonication for the purpose of lysis of E. coli cells may not be viable for scale-up.
However, this variable would need to be tested for complete verification.
4.4 Synergistic lysis using lysozyme and Triton X-100
A series of experiments was performed to determine the synergistic capabilities of using
both lysozymes and Triton X-100 to lyse E. coli cells. It is hypothesized that the utilization of these
two chemicals will provide a larger level of lysis than what would be obtained if each method was
used separately. Lysozymes break down the polysaccharide chains of peptidoglycan which
surround the inner membrane of the E. coli cell [44], [49]–[52]. Lysozyme achieves this disruption
by attacking the β-1,4 linkages of the polysaccharide chains [44], [49]–[52]. However, due to the
fact that gram-negative bacteria also have an outer membrane surrounding the peptidoglycan layer,
it is more difficult for lysozymes to reach the peptidoglycan layer to achieve complete lysis. Triton
X-100 is a non-ionic detergent that is able to solubilize the outer and inner membranes of the E.
coli cell [37], [38]. However, this is only the case when there are no divalent cations present in the
lysis buffer [41]. Therefore, the utilization of Triton X-100 for the solubilisation of the inner and
outer membrane coupled with lysozymes and their ability to break down the peptidoglycan layer
leads to a higher amount of cellular lysis. Note that for these experiments, 14-2M E. coli bacteria
was used instead of 2-3P E. coli. This was due to the fact that the production of the company
79
changed in the course of the project. The antibody produced, 14-2M, is a c.a. 16 kDa monomeric
form of the antibody of interest which localizes in the periplasmic space of the E. coli cell; 2-3P is
c.a. 25 kDa pentameric form of the antibody which localizes in the intracellular space.
4.4.1 Results and discussion
Figure 4-19 provides a comparison of SDS-PAGE gels where Gel A represents proteins
from E. coli cells that have been lysed solely using lysozymes while Gel B represents proteins
from E. coli cells that have been lysed using both lysozymes and Triton X-100. The “total”,
“soluble” and “pellet” samples were run on the gel which are denoted in total (T), soluble (S) and
insoluble (P).
Figure 4-19: SDS-PAGE of E. coli proteins. L denotes “protein ladder”, SD denotes a BSA standard, PP denotes the
purified protein of interest, T denotes total proteins, S denotes soluble proteins, P denotes insoluble proteins.
Numerical values denote the cell lysate sample. A: Cells lysed solely in the presence of lysozyme (1350 U per g of
wet cell paste). B: Cells lysed in the presence of both lysozymes (1350 U per g of wet cell paste) and Triton X-100 1
% v/v final concentration.
Comparisons were then made between the “total” and “soluble” samples in order to prepare
the histogram of Figure 4-20. The visual inspection of each gel reveals that proteins found in each
soluble lane are closer in intensity to their corresponding “total lane” in Gel B where synergistic
lysis occurred. This observation indicates that more cellular lysis occurs under the conditions
performed for the synergistic lysis (Gel B) than with lysis using only lysozymes (Gel A). It can
80
also be seen from Figure 4-19 that fainter bands are found in the pellet (P) lanes which is a clear
indication that a higher amount of lysis and solubilisation occurred. Solubilisation of proteins can
also be attributed to the fact that Triton-X 100 has been reported to be able to solubilize proteins
[37], [38]. Quantitative analysis of the gels from Figure 4-19 can be observed in Figure 4-20.
Figure 4-20: A comparison of the amount of fractional lysis of E. coli cells using either solely lysozymes (1350 U per
g of wet cell paste) or a combination of lysozymes (1350 U per g of wet cell paste) and Triton X-100 1 % v/v final
concentration. Samples were run on SDS-PAGE and evaluated through densitometry using ImageJ software to
determine the fractional amount of lysis performed. Fractional conversion of lysis was determined by dividing the
amount of proteins present in the “soluble” lanes in each gel by the amount of proteins present in each corresponding
“total” lane.
Figure 4-20 shows that the amount of cellular lysis obtained from the synergistic lysis is
substantially larger than the amount of cellular lysis obtained using only lysozymes with an
increase in lysis of approximately 35% (0.71 0.07, 0.36 0.03 for enzymatic lysis with and
without Triton X-100 respectively). As stated earlier, Triton X-100 solubilizes the outer membrane
of gram-negative cells while lysozymes break apart the peptidoglycan layer. After both the outer
membrane and the peptidoglycan layer have been removed from the cell, the cell is then able to
lyse and release its intracellular proteins. If only lysozymes were used, the outer membrane would
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
1
Fra
ctio
nal
Am
ount
of
Lysi
s
Lysozyme Lysozyme + Triton X-100
81
remain intact and the lysozymes would be unable to attack the peptidoglycan layer thereby limiting
the amount of lysis.
4.4.2 Summary
Results have shown that lysing E. coli cells in the presence of both lysozymes and Triton
X-100 results in 50 % more lysis compared to lysing only with lysozymes (0.71 0.07, 0.36
0.03 for enzymatic lysis with and without Triton X-100 respectively). An experiment where E.
coli lysis based solely on exposure to Triton X-100 could be performed to better elucidate the
effect of Triton X-100. Future experimentation would involve testing multiple concentrations of
Triton X-100 or lysozyme to determine an optimal formula for lysis.
4.5 Overall Comparison
Since nearly all proteins can be released using a high pressure homogenizer, results
obtained with the high pressure homogenizer were used to demonstrate the amount of lysis that is
achievable and for setting a benchmark for comparison between the other forms of lysis used in
this investigation. This technique displayed that it is possible to achieve maximum lysis and protein
release. However, a high pressure homogenizer can be an expensive piece of equipment for a start-
up company with a capital cost of over $22 000. The results of different lysis methods were
presented and evaluated to determine which one would provide a cheaper alternative to high
pressure homogenization to process E. coli cells. The methods that were investigated were bead
milling, sonication and chemical/enzymatic lysis.
The experiments involving the bead milling technique did not produce the same amount of
lysis that was seen with high pressure homogenization. However, under the optimal conditions
that were tested, a very positive fractional lysis of 0.79 was observed (79 % of the intracellular
82
proteins were released in the soluble fraction). Furthermore, it was demonstrated that cheaper glass
beads were available for lysis that did not reduce efficiency. From a capital cost perspective bead
milling is a cheaper alternative compared to high pressure homogenization ($700 for the unit used
during experimentation with the potential to lyse 175 mL of cell suspension). Finally, there is
potential to design and build a large bead milling unit in-house since highly specialized materials
are not required (as opposed to building a high pressure homogenizer or a sonicator).
For sonication, the objective of these experiments was to rapidly screen for the most
appropriate conditions that would lyse the E. coli cells and release large amounts of the specific
protein. Compared to other methods, it is clear that sonication does not produce a high level of
lysis. When comparing these results with those of bead milling, it can be seen that bead milling
generates higher amounts of fractional lysis (0.79 compared to 0.63 which was the highest
fractional lysis seen with each technique). Furthermore, the capital costs associated with sonication
are quite high (over $4 000 per unit). Therefore, sonication was not selected as the recommended
lysis technique.
In regards to the lysozyme/Triton X-100 technique the time required to achieve over 70 %
of lysis (30-60 min) was considerably longer than the 3 min it takes to achieve a similar amount
of lysis through bead milling. This increase in time can potentially increase significantly the
antibody production time if this form of lysis were to be used. Also, the cost of purchasing
lysozymes can be a deterrent for opting for chemical lysis ($11.33 per g of 9000 U/mg enzyme).
If 1350 U are required to lyse 100 g of cell paste (as described above) the cost to lyse 1 kg of wet
cell paste would be $1.70. This operating cost may prove to be unfeasible if the amount of protein
produced per kg of cells is determined to be unfavourable. The power input for a bead milling
process can be considered to be the major operating cost associated with that form of lysis which
83
would be much less than for chemical lysis. It should be noted that there would also be some labour
costs associated with the bead milling technique. These costs would predominately be due to
maintenance of the apparatus and cleaning/replacement of the glass beads. However, since bead
milling can be designed in a continuous configuration, these would not be day-to-day costs but
rather only seen when shutdown is required. It has been determined that lysing with lysozymes
(and Triton X-100) will not be continued based upon the relatively large amount of time required
to lyse the cells and the large operating costs associated with the cost of the lysozymes.
Therefore, the lysis technique that has been selected for the purpose of purifying the target
protein of interest is bead milling.
84
CHAPTER 5: CONCLUSION AND
RECOMMENDATIONS
5.1 Conclusion
The purpose of this research was to make a recommendation to a local start-up company,
Ab Biotech Inc., for the selection of the lysis technique for E. coli cells for the release a cytoplasmic
sdAb. Multiple lysis techniques were tested throughout this thesis including high pressure
homogenization, bead milling, sonication, and synergistic lysis with lysozymes and Triton X-100.
Since high pressure homogenization is considered to be an industry standard (almost 100 % lysis
achieved after a single pass), this form of lysis was used as a benchmark for comparison. Based
on the analysis presented, it was recommended that Ab Biotech Inc. implement bead milling for
their future large-scale operations. A high extent of fractional lysis (0.7 0.10) was achieved over
a short period of lysis time (3 min with 0.3 mm beads). Also, there are many parameters that could
be optimized in order to further reduce costs. One example of this was the choice to recommend a
0.3 mm glass bead mixture which could was considerably cheaper ($9.50/kg) compared to 0.1 mm
or 0.5 mm uniform bead sizes ($79.2/kg).
5.2 Recommendations
The next logical step would be to see how well bead milling can lyse E. coli cells in a
continuous bead milling process. It would be pertinent to demonstrate that at least 0.70 fractional
lysis could be obtained under 3 min residence time under a continuous configuration with a 0.3
mm mixture of glass beads. A small-scale continuous bead milling unit could be designed and
tested prior to designing a large-scale unit. Important parameters to consider when designing a
85
continuous bead milling unit would be the volume, lysate flow rate and the system to separate the
glass beads from the lysate.
Another task that would need to be completed would be an extensive economic analysis.
Unfortunately, throughout the entirety of the investigation there was very little expression of the
target protein. Therefore, it was very difficult to perform a suitable economic analysis since the
profitability of the entire process (from fermentation to purification) is contingent on a high yield
of the target protein. Once a sufficient yield is achieved, then a proper economic analysis can be
performed.
86
CHAPTER 6: REFERENCES
[1] R. Lyng, “Gender Preselection in Farm Animals,” Health Guidence, 2014. [Online].
Available: http://www.healthguidance.org/entry/8178/1/Gender-Preselection-in-Farm-
Animals.html. [Accessed: 31-Mar-2016].
[2] G. E. Seidel, “Overview of sexing sperm,” Theriogenology, vol. 68, no. 3, pp. 443–446,
2007.
[3] K. Hamano, “Sex preselection in bovine by separation of X- and Y-chromosome bearing
spermatozoa,” J Reprod Dev, vol. 53, no. 1, pp. 27–38, 2007.
[4] K. H. Lu, D. G. Cran, and G. E. Seidel, “In vitro fertilization with flow-cytometrically-
sorted bovine sperm,” Theriogenology, vol. 52, no. 8, pp. 1393–1405, 1999.
[5] J. M. Morrell and H. Rodriguez-Martinez, “Practical applications of sperm selection
techniques as a tool for improving reproductive efficiency.,” Vet. Med. Int., vol. 2011, 2010.
[6] D. Rath, S. Barcikowski, S. De Graaf, W. Garrels, R. Grossfeld, S. Klein, W. Knabe, C.
Knorr, W. Kues, H. Meyer, J. Michl, G. Moench-Tegeder, C. Rehbock, U. Taylor, and S.
Washausen, “Sex selection of sperm in farm animals: Status report and developmental
prospects,” Reproduction, vol. 145, no. 1, 2013.
[7] D. Rath and L. A. Johnson, “Application and Commercialization of Flow Cytometrically
Sex-Sorted Semen,” Reprod. Domest. Anim., vol. 43, no. 2, pp. 338–346, 2008.
[8] J. L. Schenk, D. G. Cran, R. W. Everett, and G. E. Seidel, “Pregnancy rates in heifers and
cows with cryopreserved sexed sperm: Effects of sperm numbers per inseminate, sorting
pressure and sperm storage before sorting,” Theriogenology, vol. 71, no. 5, pp. 717–728,
2009.
[9] D. Filpula, “Antibody engineering and modification technologies,” Biomol. Eng., vol. 24,
no. 2, pp. 201–215, 2007.
[10] M. Arbabi-Ghahroudi, R. MacKenzie, and J. Tanha, “Selection of non-aggregating VH
binders from synthetic VH phage-display libraries,” Methods Mol Biol, vol. 525, pp. 187–
216, xiii, 2009.
[11] M. L. Shuler and F. Kargi, Bioprocess Engineering Basic Concepts, Second. Upper Saddle
River, New Jersey: Prentice Hall, 2002.
[12] J. C. Sharpe and K. M. Evans, “Advances in flow cytometry for sperm sexing,”
Theriogenology, vol. 71, no. 1, pp. 4–10, 2009.
[13] G. R. Welch and L. A. Johnson, “Sex preselection: Laboratory validation of the sperm sex
ratio of flow sorted X- and Y-sperm by sort reanalysis for DNA,” Theriogenology, vol. 52,
no. 8, pp. 1343–1352, 1999.
[14] D. L. Garner and G. E. Seidel, “History of commercializing sexed semen for cattle,”
Theriogenology, vol. 69, no. 7, pp. 886–895, 2008.
87
[15] L. A. Johnson, J. P. Flook, M. V Look, and D. Pinkel, “Flow sorting of X and Y
chromosome-bearing spermatozoa into two populations.,” Gamete Res., vol. 16, no. 1, pp.
1–9, 1987.
[16] L. A. Johnson and G. R. Welch, “Sex preselection: High-speed flow cytometric sorting of
X and Y sperm for maximum efficiency,” Theriogenology, vol. 52, no. 8, pp. 1323–1341,
1999.
[17] T. K. Suh, J. L. Schenk, and G. E. Seidel, “High pressure flow cytometric sorting damages
sperm,” Theriogenology, vol. 64, no. 5, pp. 1035–1048, 2005.
[18] G. E. Seidel, “Economics of selecting for sex: The most important genetic trait,”
Theriogenology, vol. 59, no. 2, pp. 585–598, 2003.
[19] D. L. Garner, “Flow cytometric sexing of mammalian sperm,” Theriogenology, vol. 65, no.
5 SPEC. ISS., pp. 943–957, 2006.
[20] L. A. Johnson, “Sexing mammalian sperm for production of offspring: The state-of-the-
art,” Anim. Reprod. Sci., vol. 60–61, pp. 93–107, 2000.
[21] S. R. Blecher, R. Howie, S. Li, J. Detmar, and L. M. Blahut, “A new approach to
immunological sexing of sperm,” Theriogenology, vol. 52, no. 8, pp. 1309–1321, 1999.
[22] P. Holliger and P. J. Hudson, “Engineered antibody fragments and the rise of single
domains.,” Nat. Biotechnol., vol. 23, no. 9, pp. 1126–1136, 2005.
[23] W. W. Gibbs, “Nanobodies,” Sci. Am., vol. 293, no. 2, pp. 78–83, 2005.
[24] A. Bell, Z. J. Wang, M. Arbabi-Ghahroudi, T. a. Chang, Y. Durocher, U. Trojahn, J.
Baardsnes, M. L. Jaramillo, S. Li, T. N. Baral, M. O’Connor-McCourt, R. MacKenzie, and
J. Zhang, “Differential tumor-targeting abilities of three single-domain antibody formats,”
Cancer Lett., vol. 289, no. 1, pp. 81–90, 2010.
[25] M. Arbabi-Ghahroudi, J. Tanha, and R. Mackenzie, “Prokaryotic expression of antibodies
and affibodies,” pp. 501–519, 2005.
[26] E. R. Goldman, G. P. Andersen, J. L. Liu, J. B. Delehanty, L. J. Sherwood, L. E. Osborn,
L. B. Cummins, and A. Hayhurst, “Facile generation of heat-stable antiviral and antitoxin
single domain antibodies from a semisynthetic llama library,” Anal. Chem., vol. 78, no. 24,
pp. 8245–8255, 2006.
[27] M. M. Harmsen and H. J. De Haard, “Properties, production, and applications of camelid
single-domain antibody fragments,” Appl. Microbiol. Biotechnol., vol. 77, no. 1, pp. 13–22,
2007.
[28] S. Muyldermans, “Single domain camel antibodies: Current status,” Rev. Mol. Biotechnol.,
vol. 74, no. 4, pp. 277–302, 2001.
[29] D. Saerens, F. Frederix, G. Reekmans, K. Conrath, K. Jans, L. Brys, L. Huang, E. Bosmans,
G. Maes, G. Borghs, and S. Muyldermans, “Engineering camel single-domain antibodies
and immobilization chemistry for human prostate-specific antigen sensing,” Anal. Chem.,
vol. 77, no. 23, pp. 7547–7555, 2005.
88
[30] J. Zhang, Q. Li, T. D. Nguyen, T. L. Tremblay, E. Stone, R. To, J. Kelly, and C. Roger
MacKenzie, “A pentavalent single-domain antibody approach to tumor antigen discovery
and the development of novel proteomics reagents,” J. Mol. Biol., vol. 341, no. 1, pp. 161–
169, 2004.
[31] C. S. Stewart, C. R. MacKenzie, and J. Christopher Hall, “Isolation, characterization and
pentamerization of alpha-cobrotoxin specific single-domain antibodies from a naïve phage
display library: Preliminary findings for antivenom development,” Toxicon, vol. 49, no. 5,
pp. 699–709, 2007.
[32] E. Stone, T. Hirama, J. Tanha, H. Tong-Sevinc, S. Li, C. R. MacKenzie, and J. Zhang, “The
assembly of single domain antibodies into bispecific decavalent molecules”, J. Immunol.
Methods, vol. 318, no. 1–2. 2007.
[33] J. Zhang, J. Tanha, T. Hirama, N. H. Khieu, R. To, H. Tong-Sevinc, E. Stone, J. R. Brisson,
and C. R. MacKenzie, “Pentamerization of single-domain antibodies from phage libraries:
A novel strategy for the rapid generation of high-avidity antibody reagents,” J. Mol. Biol.,
vol. 335, no. 1, pp. 49–56, 2004.
[34] N. Ruiz, D. Kahne, and T. J. Silhavy, “Advances in understanding bacterial outer-membrane
biogenesis.,” Nat. Rev. Microbiol., vol. 4, no. 1, pp. 57–66, 2006.
[35] Encyclopedia of Microbiology. Academic Press, 2009.
[36] T. R. Hopkins, “Physical and chemical cell disruption for the recovery of intracellular
proteins,” Bioprocess Technol., no. 12, pp. 57–83, 1991.
[37] C. A. Schnaitman, “Solubilization of the cytoplasmic membrane of Escherichia coli by
Triton X-100.,” J. Bacteriol., vol. 108, no. 1, pp. 545–552, 1971.
[38] D. Hettwer and H. Wang, “Protein release from Escherichia coli cells permeabilized with
guanidine-HCl and Triton X100.,” Biotechnol. Bioeng., vol. 33, no. 7, pp. 886–895, 1989.
[39] S. T. L. Harrison, “Bacterial Cell Disruption: A Key Unit Operation in the Recovery of
Intracellular Products,” Biotechnol. Adv., vol. 9, pp. 217–240, 1991.
[40] C. Filip, G. Fletcher, J. L. Wulff, and C. F. Earhart, “Solubilization of the cytoplasmic
membrane of Escherichia coli by the ionic detergent sodium-lauryl sarcosinate.,” J.
Bacteriol., vol. 115, no. 3, pp. 717–22, 1973.
[41] M. L. DePamphilis, “Dissociation and reassembly of Escherichia coli outer membrane and
of lipopolysaccharide, and their reassembly onto flagellar basal bodies,” J. Bacteriol., vol.
105, no. 3, pp. 1184–1199, 1971.
[42] H. Umakoshi, Y. Fukuta, and R. Kuboi, “Utilization of cell response under heat, chemical,
and combined stresses for selective recovery of cytoplasmic beta-galactosidase from
Escherichia coli cells,” Biotechnol Prog, vol. 14, no. 6, pp. 909–912, 1998.
[43] C. A. Schnaitman, “Effect of ethylenediaminetetraacetic acid, triton x-100, and lysozyme
on the morphology and chemical composition of isolated cell walls of Escherichia coli,” J
Bacteriol, vol. 108, no. 1, pp. 553–563, 1971.
89
[44] D. S. Feingold, J. N. Goldman, and H. M. Kuritz, “Locus of the Action of Serum and the
Role of Lysozyme in the Serum Bactericidal Reaction,” J. Bacteriol, vol. 96, no. 6, pp.
2118–2126, 1968.
[45] J. K. Branen and P. M. Davidson, “Enhancement of nisin, lysozyme, and monolaurin
antimicrobial activities by ethylenediaminetetraacetic acid and lactoferrin,” Int. J. Food
Microbiol., vol. 90, pp. 63–74, 2004.
[46] R. J. Falconer, B. K. O’Neill, and A. P. J. Middelberg, “Chemical treatment of Escherichia
coli: 1. Extraction of intracellular protein from uninduced cells,” Biotechnol. Bioeng., vol.
53, no. 5, pp. 453–458, 1997.
[47] H. Brown, Chemistry of the Cell Interface, Part 2. Elsevier, 2012.
[48] A. Pellegrini, U. Thomas, R. von Fellenberg, and P. Wild, “Bactericidal activities of
lysozyme and aprotinin against Gram-negative and Gram-positive bacteria related to their
basic character,” J. Appl. Bacteriol., vol. 72, no. 3, pp. 180–187, 1992.
[49] P. Jolles, “Recent Developments in the Study of Lysozymes.,” Angew. Chem. Int. Ed. Engl.,
vol. 3, no. 1, pp. 28–36, 1964.
[50] C. R. Dean and O. P. Ward, “The use of EDTA or Polymyxin with Lysozyme for the
Recovery of Intracellular Products from Escherichia coli,” Biotechnol. Tech., vol. 6, no. 2,
pp. 133–138, 1992.
[51] P. Wild, A. Gabrieli, E. M. Schraner, A. Pellegrini, U. Thomas, P. M. Frederik, M. C.
Stuart, and R. Von Fellenberg, “Reevaluation of the effect of lysoyzme on Escherichia coli
employing ultrarapid freezing followed by cryoelectronmicroscopy or freeze substitution.,”
Microsc. Res. Tech., vol. 39, no. 3, pp. 297–304, 1997.
[52] A. J. Kirby, “The lysozyme mechanism sorted -- after 50 years.,” Nat. Struct. Biol., vol. 8,
no. 9, pp. 737–739, 2001.
[53] R. Repaske, “Lysis of Gram-Negative Organisms and the Role of Versene,” Biochim.
Biophys. Acta, vol. 30, pp. 225–232, 1958.
[54] I. N. Cell and W. Rigidity, “Lysozyme sensitivity of the cell wall of Pseudomonas
aeruginosa: further evidence for the role of the non-peptidoglycan components in cell wall
rigidity,” vol. 1, no. 2, 1966.
[55] E. Y. Wuytack, A. M. J. Diels, and C. W. Michiels, “Bacterial inactivation by high-pressure
homogenisation and high hydrostatic pressure,” Int. J. Food Microbiol., vol. 77, no. 3, pp.
205–212, 2002.
[56] D. Kopeliovich, “High Pressure Homogenization,” SubsTech Substances & Technologies,
2013. Available at http://www.substech.com/dokuwiki/doku.php?id=homogenization.
[Accessed: 12-Jul-2016]. Copyright information available at:
https://creativecommons.org/licenses/by-nc-sa/3.0/ [Accessed: 12-Jul-2016].
90
[57] R. N. Ramanan, T. Tey, T. C. Ling, A. B. Ariff, and U. P. M. Serdang, “Classification of
Pressure Range Based on the Characterization of Escherichia coli Cell Disruption in High
Pressure Homogenizer”, Institute of Bioscience , University Putra Malaysia , 43400
Serdang, Selangor, Malaysia Department of Chemical and Environmental Engineering, Am.
J. Biochem. Biotechnol., vol. 5, no. 1, pp. 21–29, 2009.
[58] L. P. Fonseca and J. M. S. Cabral, “Penicillin acylase release from Escherichia coli cells by
mechanical cell disruption and permeabilization,” J. Chem. Technol. Biotechnol., vol. 77,
no. 2, pp. 159–167, 2002.
[59] A. M. J. Diels, L. Callewaert, E. Y. Wuytack, B. Masschalck, and C. W. Michiels,
“Moderate Temperatures Affect Escherichia coli Inactivation by High‐Pressure
Homogenization Only through Fluid Viscosity,” Biotechnol. Prog., vol. 20, no. 5, pp. 1512–
1517, 2004.
[60] A. M. J. Diels, L. Callewaert, E. Y. Wuytack, B. Masschalck, and C. W. Michiels,
“Inactivation of Escherichia coli by high-pressure homogenisation is influenced by fluid
viscosity but not by water activity and product composition.,” Int. J. Food Microbiol., vol.
101, no. 3, pp. 281–91, 2005.
[61] H. Anand, B. Balasundaram, A. B. Pandit, and S. T. L. Harrison, “The effect of chemical
pretreatment combined with mechanical disruption on the extent of disruption and release
of intracellular protein from E. coli,” Biochem. Eng. J., vol. 35, no. 2, pp. 166–173, 2007.
[62] R. B. Brown and J. Audet, “Current techniques for single-cell lysis,” J. R. Soc. Interface,
vol. 5, no. 2, pp. S131–S138, 2008.
[63] D. Kopeliovich, “Ultrasonic homogenization,” SubsTech Subst. Technol. Available at
http://www.substech.com/dokuwiki/doku.php?id=homogenization. [Accessed: 12-Jul-
2016]. Copyright information available at: https://creativecommons.org/licenses/by-nc-
sa/3.0/ [Accessed: 12-Jul-2016].
[64] C. W. Ho, T. K. Chew, T. C. Ling, S. Kamaruddin, W. S. Tan, and B. T. Tey, “Efficient
mechanical cell disruption of Escherichia coli by an ultrasonicator and recovery of
intracellular hepatitis B core antigen,” Process Biochem., vol. 41, no. 8, pp. 1829–1834,
2006.
[65] E. M. Fykse, J. S. Olsen, and G. Skogan, “Application of sonication to release DNA from
Bacillus cereus for quantitative detection by real-time PCR,” J. Microbiol. Methods, vol.
55, no. 1, pp. 1–10, 2003.
[66] I. Hua and J. E. Thompson, “Inactivation of Escherichia coli by sonication at discrete
ultrasonic frequencies,” Water Res., vol. 34, no. 15, pp. 3888–3893, 2000.
[67] J. X. Feliu, R. Cubarsi, and A. Villaverde, “Optimized release of recombinant proteins by
ultrasonication of E. coli cells.,” Biotechnol. Bioeng., vol. 58, no. 5, pp. 536–40, 1998.
[68] S. Dakubu, “Cell Inactivation by Ultrasound,” vol. XVIII, pp. 465–471, 1976.
[69] R. S. Singh, “A comparative study on cell disruption methods for release of aspartase from
E. coli K-12,” Indian J. Exp. Biol., vol. 51, no. 11, pp. 997–1003, 2013.
91
[70] A. Villaverde and J. X. Feliu, “An optimized ultrasonication protocol for bacterial cell
disruption and recovery of galactosidase fusion proteins,” Biotechnol. Tech., vol. 10, no. 7,
pp. 713–716, 1995.
[71] L. Benov and J. Al-Ibraheem, “Disrupting Escherichia coli: A Comparison of Methods,” J.
Biochem. Mol. Biol., vol. 35, no. 4, pp. 428–431, 2002.
[72] M. Mojsin, G. Nikcevic, N. Kovacevic-Grujicic, T. Savic, I. Petrovic, and M. Stevanovic,
“Purification and functional analysis of the recombinant protein isolated from E. coli by
employing three different methods of bacterial lysis,” J. Serbian Chem. Soc., vol. 70, no. 7,
pp. 943–950, 2005.
[73] D. E. Handley, Systems Approach to Analyzing the TGFBeta/SMAD3 Gene Regulatory
Pathway in A549 Cells. Proquest, 2008.
[74] C. W. Ho, W. S. Tan, S. Kamaruddin, T. C. Ling, and B. T. Tey, “The direct recovery of
recombinant hepatitis B core antigen from disruptate derived from continuous-flow bead
milling.,” Biotechnol. Appl. Biochem., vol. 50, no. Pt 1, pp. 49–59, 2008.
[75] S. Gerber, K. Kirchhof, J. Kressler, C. E. H. Schmelzer, C. Scholz, T. C. Hertel, and M.
Pietzsch, “Cloning, expression, purification, and characterization of a designer protein with
repetitive sequences,” Protein Expr. Purif., vol. 59, no. 2, pp. 203–214, 2008.
[76] H. Schütte, K. H. Kroner, H. Hustedt, and M.-R. Kula, “Experiences with a 20 litre
industrial bead mill for the disruption of microorganisms,” Enzyme Microb. Technol., vol.
5, no. 2, pp. 143–148, 1983.
[77] Andriantoangkadirjo85, “Equipments Used In Cosmetics Production,” WordPress, 2011.
[Online]. Available: https://andriantoangkadirjo85.wordpress.com/tag/bead-mill/.
[Accessed: 02-Apr-2016].
[78] A. P. J. Middleberg, “Process-Scale Disruption of Microorganisms.” Elsevier Science Inc.
[79] U. Bakir and H. Hamamci, “The effect of freeze-thawing on the release of intracellular
proteins from Escherichia coli by means of a bead mil,” World J. Microbiol. Biotechnol.,
vol. 13, no. 4, pp. 475 – 477, 1997.
[80] R. F. Itzhaki and D. M. Gill, “A micro-biuret method for estimating protein,” Anal.
Biochem, no. 9, pp. 401–410, 1964.
[81] R. K. Scopes, Protein Purification: Princliples and Practice. Springer Science and Business
Media, 1982.
[82] O. Warburg and W. Christian, “Isolierung und Kristallisation des Garungsferments
Enolase,” Biochem. Z., no. 310, pp. 384–421, 1941.
[83] R. K. Scopes, “Measurement of protein spectrophotometry at 205 nm,” Anal. Biochem, no.
59, pp. 277–282, 1974.
[84] M. M. Bradford, “A rapid and sensitive method for the quantification of microgram
quantities of protein utilizing the principle of protein-dye binding.,” Anal. Biochem, no. 72,
pp. 248–254, 1976.
92
[85] T. Spector, “Refinement of the Coomassie blue method of protein quantitation,” Anal.
Biochem, vol. 86, pp. 142–146, 1978.
[86] M. De Mey, G. J. Lequeux, J. Maertens, C. I. De Muynck, W. K. Soetaert, and E. J.
Vandamme, “Comparison of protein quantification and extraction methods suitable for E.
coli cultures,” Biologicals, vol. 36, no. 3, pp. 198–202, 2008.
[87] R. J. Randall and A. Lewis, “Protein Measurement with the Folin phenol reagent,” J. Biol.
Chem., vol. 193, no. 1, pp. 265–275, 1951.
[88] O. Folin and V. Ciocalteau, “Tyrosine and tryptophan determination in proteins,” J. Biol.
Chem., no. 73, pp. 627–650, 1927.
[89] G. L. Peterson, “Review of the Folin phenol protein quantitation method of Lowry,
Rosebrough, Farr, and Randall,” Anal. Biochem, no. 100, pp. 201–220, 1979.
[90] K. J. Wiechelman, R. D. Braun, J. D. Fitzpatrick, K. Smith, R. I. Krohn, G. T. Hermanson,
A. K. Mallia, F. H. Gartner, M. D. Provenzano, E. K. Fujimoto, N. M. Goeke, B. J. Olson,
and D. C. Klenk, “Investigation of the Bicinchoninic Acid Protein Assay: Identification of
the Groups Responsible for Color Formation,” Analytical biochemistry, vol. 237, pp. 231–
237, 1988.
[91] J. M. Walker, The Protein Protocols Handbook, 1st ed. Totowa, New Jersey: Humana Press
Inc., 1996.
[92] A. Singh, Enzyme Assays. Regency Publication, 2007.
[93] R. B. Corley, A Guide to Methods In the Biomedical Sciences. Springer Science & Business
Media, 2004.
[94] M. D. Abramoff, P. J. Magalhaes, and S. J. Ram, “Image processing with imageJ,”
Biophotonics Int., vol. 11, no. 7, pp. 36–41, 2004.
[95] D. J. King, Applications And Engineering of Monoclonal Antibodies. CRC Press, 1998.
[96] J. A. Bornhorst and J. J. Falke, “Purification of Proteins Unsing Polyhistidine Affinity
Tags,” Methods Enzym., no. 326, pp. 245–254, 2000.
[97] F. W. Studier, “Protein production by auto-induction in high density shaking cultures.,”
Protein Expr. Purif., vol. 41, no. 1, pp. 207–234, 2005.
[98] R. Milo and R. Phillips, Cell Biology by the Numbers. Garland Science, 2015.
[99] “APV Rannie and Gaulin Homogenizers,” SPX Flow Technology, 2013. [Online].
Available: http://pdf.directindustry.com/pdf/apv/apv-homogenizers/5697-394007.html.
[100] BioSpec., “BeadBeater Operating Instructions.” [Online]. Available:
http://www.biospec.com/instructions/beadbeater/. [Accessed: 09-Apr-2016].
[101] S. T. Harrison, J. S. Daniel, and H. A. Chase, “The effect of culture history on the disruption
of Alcaligenes eutrophus by high pressure homogenisation,” in Recovery of Biological
Products V, 1990, pp. 13–18.
[102] H. Felix, “Permeabilised cells,” Anal. Biochem., no. 120, pp. 211–234, 1982.
93
[103] N. M. Fish and M. D. Lilly, “The Interactions Between Fermentation and Protein
Recovery,” Biotechnology, vol. 2, no. 7, pp. 623–627, 1984.
[104] I. G. Golovina, N. D. Saveleva, L. G. Loginova, and G. A. Zavarzin, “Lysis of
Hydrogenomonas eutropha Z-1 cells by the enzymes of Micromonospora vulgaris,”
Mikrobiologiya, vol. 5, no. 42, pp. 899–903, 1973.
[105] R. H. Cumming, J. Tuffnell, and G. Street, “The Dependence of Protein Release from
Bacillus amyloliquefaciens on the Growth Phase in Batch Culture Growth of Cells,” no.
Table I, pp. 6–8, 1984.
[106] Y. S. Lee, S. C. Park, A. L. Goldber, and C. H. Chung, “Protease So from Escherichia coli
preferentially degrades oxidatively damaged glutamine synthetase,” J. Biol. Chem., vol.
263, no. 14, pp. 6643–6646, 1988.